Graphical Abstract Figure
Graphical Abstract Figure
Close modal

Abstract

Recently, an ejector refrigeration system (ERS) has been a promising cooling strategy with waste heat utilization and minimization of power consumption by evading the compressor. However, analyzing the intricate flow structure inside the ejector and the corresponding coefficient of performance enhancement are major challenges of an ERS. The type of working fluid, design specifications, and working conditions significantly affect the ejector behavior. The environmental issues caused by the leakage of the most popular high-GWP refrigerant R134a divulge the need for low-GWP alternatives. Moreover, the effect of critical design specifications such as area ratio (AR) and nozzle exit position (NXP) for these alternatives is not explored yet. Therefore, five low-GWP alternates for R134a, namely R1234yf, R1243zf, R152a, R513a, and R440a, are tested numerically under wide-ranging operating conditions. In addition, the ejector performance for all refrigerants is examined for seven distinct ARs and five different NXPs. The effect of the refrigerant variant and NXP on the internal flow structures of the ejector is also analyzed. Besides, the study is extended to find the optimal NXP at various operating temperatures using R1234yf refrigerant. In most cases, the higher entrainment ratio (ER) is obtained with R1234yf and R1243zf, and the increase in AR has a positive effect on the ER. The impact of the NXP is higher at condenser temperature with minimal waste heat in the generator. Irrespective of the operating conditions for R1234yf, the optimum NXP is obtained as 10 mm, which is 1.67 times the constant-area mixing chamber diameter.

1 Introduction

The waste heat from automobile exhausts, industrial processes, and air-conditioning facilities is tending toward global warming and ozone layer depletion. For example, a typical vapor compression refrigeration system (VCRS) used to provide air-conditioning in automobiles increases fuel consumption by up to 12–17%. In addition, the burning of fossil fuels in automobiles emits exhaust gases that cause long-term damage to the environment. In addition, the energy wasted through exhaust gases, coolant, and engine compartments of any automobile tolls up to 65% of total consumption [1,2]. Therefore, the utilization of this waste heat to operate a cooling system is an encouraging way of energy management with reduced fuel consumption. There are a few promising technologies that provide a cooling effect by recovering the process heat, such as absorption, adsorption, desiccant, and ejector refrigeration systems (ERS). Among these, the ERS is drawing more attention in recent days due to its simple and compact structure with cooling capacity from a few Watts to a kilowatts range, economic viability, and ease of maintenance, in spite of a lower coefficient of performance (COP) in the range of 0.1–0.4. However, the adsorption, absorption, and desiccant systems are not suitable for small-range applications like automotive refrigeration [3,4]. Therefore, further investigation of an ERS-based cooling system is needed to understand the complex flow dynamics across the ejector toward superior cooling capacity and COP.

To improve the COP of an ERS, several investigations are performed by varying the operating parameters, design specifications, and refrigerant types [5]. These studies indicate that condenser, evaporator, and generator temperatures are the key attributes significantly influencing primary and secondary flowrates within the system. In parallel, area ratio (AR), primary nozzle exit position (NXP), length and diameter of constant-area mixing section, convergence angle of constant-pressure mixing section, and divergence angle of the diffuser are other important geometric parameters that affect the overall ERS performance [6,7]. Efficient entrainment of secondary flow into the ejector necessitates a minimum primary flow pressure and optimal nozzle Mach number and ejector area ratio [8]. Although the ERS performance is predominantly affected by AR, the NXP also significantly impacts ejector behavior [9]. Commonly, the analytical model of an ejector is configured based on two approaches: constant-pressure mixing (CPM) and constant-area mixing (CAM), depending on the location of the NXP at the suction chamber or the constant-area section, respectively. However, the CPM superimposes the CAM modeling approach due to its utmost performance, ease of solution, and accuracy [10]. According to the ESDU design guide, the optimum NXP is 0.5–1.0 times the mixing chamber diameter, and the change in NXP can give a 100% rise in COP and cooling capacity [11]. Practically, it is difficult to specify a single optimum value for NXP due to complex flow structures and fast dynamics of working parameters [12,13]. For instance, the optimal NXP for refrigerant R141b is observed to be 1.5 times and 1.7–3 times the mixing chamber diameter in two separate investigations [14,15].

Most of the aforementioned studies are based on the 1D-CPM analogy. According to this model, at a specific distance from the NXP, the secondary flow becomes choked and begins to merge with the primary flow at constant pressure [16,17]. However, experiments confirm that the mixing phenomena inside the ejector are not simplistic [18,19]. Irreversible losses caused by an oblique shock wave from the primary nozzle are not considered by 1D models, demonstrating that the 1D-CPM theory with a single shock wave in the mixing section cannot be applied satisfactorily to all physical conditions [20]. Therefore, knowledge of the inner dynamics of this complex flow phenomenon is required to understand the behaviour working fluids, geometrical parameters, and working conditions in the mixing process. The computational fluid dynamics (CFD) technique is one such tool for visually evaluating the ejector flow field. A CFD simulation study employing six turbulence models and air as the working fluid for an ejector demonstrates excellent agreement with experimental results [21]. As air is not suitable for cooling applications, later studies focused on hydrochlorofluorocarbon (HCFC) and hydrofluorocarbon (HFC) refrigerants. CFD investigations of ERS using steam [22] and refrigerants R141b [23], R142b [20], and R1234yf [9] show that the secondary flow saturation temperature, AR, and NXP enhance ER, critical backpressure, and overall performance. It reports that the suction chamber geometry and secondary flow entry velocity do not considerably affect the flow pattern of the ejector. Therefore, the 3D and axisymmetric CFD models give almost similar results for CAM and CPM steam ejectors [24]. The CFD investigations on different ejector geometries at various operating circumstances using steam [6] and refrigerant R134a [25] reveal that the primary nozzle geometry and shock circle location can substantially affect the ejector's performance. Besides, these studies disclose that the ER is more sensitive to the throat diameter, surface roughness, diverging part length, and NXP of the primary nozzle.

HFC-134a is the utmost popular refrigerant used in refrigeration and AC applications. However, the leakage of this refrigerant can directly damage the environment owing to its higher global warming potential (GWP) value. The Montreal Protocol's Kigali amendment restricts the usage of R134a refrigerants and encourages the refrigeration industries to adopt zero-ozone depletion potential (ODP) and low-GWP refrigerants [26]. In recent years, a few new refrigerants have been introduced, such as R152a, R290 [27,28], R717, R600a [29,30], R1234yf, R1234ze(E) [31,32], R1233zdE [33], R515a, R456a, and R516a [34]. Recent studies also explored nano refrigerants, which are homogeneous mixtures of R134a, R152a, and R290 with TiO2 and Al2O3 nanoparticles (0–5 wt%) [35]. While nanoparticle concentration enhances performance, it may also adversely affect ejector longevity due to intricate flow and shock wave phenomena inside the mixing chamber. Among all, the refrigerants R152a, R600a, and R290 are highly flammable, while the performance of R515a, R456a, and R1234ze(E) is very poor compared to R134a. Therefore, R1234yf and R1233zdE have garnered a lot of interest as potential replacements for the most widely used R134a refrigerant [36]. However, the use of R134a alternatives in ejectors and the impact of ejector geometry on these alternatives and their flow dynamics have not been extensively studied. Hence, the present investigation explores the qualitative performance of recently developed sustainable refrigerants and conventional refrigerants.

The preceding literature on ERS predominantly focuses on the impact of operational and geometrical factors on their performance, utilizing conventional refrigerants like R134a. However, a comprehensive understanding of ejector performance under the influence of a wide range of eco-friendly refrigerants and their intricate flow dynamics within the ejector is yet to be extensively examined. Furthermore, optimizing ejector geometry and operating conditions specifically for these environmentally friendly refrigerants remains largely unexamined. This lack of understanding regarding the optimal utilization of environmentally friendly refrigerants within ERS hinders the development of sustainable and efficient cooling solutions that align with global environmental goals. In view of this, the current work aims to examine the influence of operational and geometrical parameters on an ejector with five eco-friendly R134a alternate refrigerants, namely R1234yf, R1243zf, R152a, R513a, and R440a, having almost similar thermo-physical properties of R134a by using ansys fluent 20.0 software. The CFD model of the ejector is developed and validated with experimental data [37]. Later, investigations are conducted to compare the ER for all the chosen refrigerants under various operating conditions, such as condenser, evaporator, and generator temperatures, by adopting the specific ejector geometry from the experimental study [38]. The impact of geometrical factors, specifically AR and NXP, on the ER is thoroughly examined by evaluating seven different ARs and five distinct NXPs for each of the selected refrigerants. Further, the study is extended to find the optimal NXP in an extensive range of operating circumstances. Ultimately, an in-depth examination of the complex flow dynamics inside the ejector is carried out.

2 Methodology

2.1 Ejector Theory.

The compressor of a traditional VCRS is replaced with a superheated ejector for any ERS (Fig. 1). The typical ejector has four major parts: the primary nozzle, the suction chamber, the constant-pressure and constant-area mixing chambers, and the diffuser. First, the primary flow from the generator with higher pressure and temperature enters and expands in the primary nozzle, creates suction, and induces secondary flow from the evaporator into the suction chamber. Next, both streams combine in the mixing chamber and undergo a normal shock wave. Finally, this mixed-stream reaches condenser pressure by recovering all the kinetic energy in the diffuser. Subsequently, the mixed stream dissipates its heat at the condenser, splitting into the primary and secondary flows. The secondary flow expands in the expansion valve, and offers cooling by absorbing heat, Qe, in the evaporator, while the primary flow is pumped back to the generator and absorbs heat, Qg.

Fig. 1
Schematic illustration of an ejector refrigeration system, the sectional view of ejector and its geometrical parameters
Fig. 1
Schematic illustration of an ejector refrigeration system, the sectional view of ejector and its geometrical parameters
Close modal
The entrainment ratio (ER) is obtained to estimate the ejector performance as the ratio of secondary mass flowrate, m˙s, to primary mass flowrate, m˙p
ER=m˙s/m˙p
(1)
The ER depends on both the ejector's working and design parameters. Under the specified operating circumstances, the COP of the ERS is expressed as
(COP)ERS=QeQg+Wpump
(2)
The pump work, Wpump, is usually insignificant as compared to Qg; hence Eq. (2) becomes
(COP)ERS=QeQg=m˙s(Δhevap)m˙p(Δhgen)=ERΔhevapΔhgen
(3)

As the change in enthalpy across the generator and evaporator are almost equal, the COP of the ERS is solely dependent on the ER [39].

The AR as the ratio of the constant-area section to the primary nozzle throat area and NXP as the distance between primary nozzle exit and constant-area section have a substantial impact on the operation of ERS [7]. The AR is obtained as the function of the primary nozzle's throat diameter (Dt) and constant-area section diameter (Dam)
AR=(DamDt)2
(4)

The other geometrical parameters required for modeling the ejector are depicted in Fig. 1. Among all, the primary nozzle exit diameter (Dne) (defines the primary flow's driving force), constant-pressure mixing chamber entrance diameter (Dpm) (at this position, secondary flow start entrains into the primary flow), and diffuser diameter (Dd) are also few critical parameters that have equal importance in modeling of the ejector [40,41]. Table 1 shows the ejector dimensions used in the current dynamics model [38].

Table 1

The ejector design parameters [38]

Design parametersValue
(mm)
Primary nozzle entry diameter (Dp)13
Primary nozzle throat diameter (Dt)3
Primary nozzle exit diameter (Dne)3.8
CAM chamber diameter (Dam)6
Diffuser exit diameter (Dd)15
Secondary flow inlet diameter (Ds)15
CPM chamber entry diameter (Dpm)12
Primary nozzle length (Ln)39
Primary nozzle divergent part length (Lne)5
CAM chamber length (Lam)20
CPM chamber length (Lpm)5
Diffuser length (Ld)70
Design parametersValue
(mm)
Primary nozzle entry diameter (Dp)13
Primary nozzle throat diameter (Dt)3
Primary nozzle exit diameter (Dne)3.8
CAM chamber diameter (Dam)6
Diffuser exit diameter (Dd)15
Secondary flow inlet diameter (Ds)15
CPM chamber entry diameter (Dpm)12
Primary nozzle length (Ln)39
Primary nozzle divergent part length (Lne)5
CAM chamber length (Lam)20
CPM chamber length (Lpm)5
Diffuser length (Ld)70

Refrigerant leaks can create various environmental issues like ozone depletion and global warming. The Kigali Amendment to the Montreal Protocol emphasizes the need to find environmentally friendly alternatives for high GWP refrigerants such as R134a. Therefore, finding a suitable replacement refrigerant for ERS is critical [16,31]. In view of this, five zero-ODP and low-GWP refrigerants are examined as alternatives for R134a, and their thermo-physical properties are reported in Table 2.

Table 2

Thermodynamic and environmental properties of the selected refrigerants

ASHRAE numberODPNet GWP 100-yrNormal boiling point
(°C)
Critical temperature
(°C)
Critical pressure
(kPa)
R134a01300−26.3101.064059
R1234yf04−29.494.73382
R1243zf04−26.64103.773518
R152a0138−25113.264517
R513A0573−29.496.53766
R440A0144−25.5112.654515
ASHRAE numberODPNet GWP 100-yrNormal boiling point
(°C)
Critical temperature
(°C)
Critical pressure
(kPa)
R134a01300−26.3101.064059
R1234yf04−29.494.73382
R1243zf04−26.64103.773518
R152a0138−25113.264517
R513A0573−29.496.53766
R440A0144−25.5112.654515

2.2 Computational Fluid Dynamics Model.

CFD is a powerful technique for analyzing and visualizing fluid flow within complex geometries. Therefore, ansys fluent 20.0 is employed here to investigate the intricate flow structure within the typical ejector. The governing equations are generated to solve the supersonic compressible flow through the ejector using mass, energy, and momentum balance equations. To simplify the model without losing generality, the flow within the ejector is assumed to be steady, and the ejector's internal surface is considered to be adiabatic. The governing equations in a compact Cartesian form can be expressed as
continuity:(ρui)xi=0
(5)
energy:[ui(ρE+P)]xi=˙(αeffTxi+uiτij)
(6)
momentum:(ρujui)xj=τijxjPxi
(7)
andτij=μeff(uixj+ujxi23ukxkδij)
(8)
where ρ,u,P,T,τ,E,μeff,δij, and αeff are the density (kg/m3), velocity (m/s), pressure (Pa), temperature (K), stress tensor (Pa), total energy (J), effective dynamic viscosity (Ns/m2), Kronecker delta function, and effective thermal conductivity (W/mK), respectively.

The dense meshes are prearranged in the primary nozzle exit and mixing section regions to handle the high incline properties developed due to the shock waves. The axisymmetric 2D model is used for ejector design since the performance and wall pressure distribution in the axisymmetric and 3D models are very similar (Fig. 2) [24]. Due to less running time requirements and better prediction of the separation and recirculation of the jet flows, the realizable kɛ with scalable wall function turbulence model is chosen [24,42,43]. The SIMPLE algorithm is employed for pressure velocity coupling to solve the discretized equations. The second-order upwind is used to discretize the convective and diffusive terms for high accuracy [44]. The detailed simulation algorithm is described in Fig. 3. In the upcoming section, the ejector behavior is evaluated under various boundary conditions, with seven refrigerants and different geometrical parameters, such as AR and NXP.

Fig. 2
Quadrilateral grid structure of 2D axisymmetric ejector
Fig. 2
Quadrilateral grid structure of 2D axisymmetric ejector
Close modal
Fig. 3
Solution algorithm of the simulation process
Fig. 3
Solution algorithm of the simulation process
Close modal

2.3 Boundary Conditions and Refrigerant Properties.

Table 3 demonstrates the boundary conditions adopted for the present analytical study. The saturation pressure corresponds to the evaporator, generator, and condenser temperatures (Tevap,Tgen,andTcond) are considered as the pressures of the upstream (secondary and primary flows) and the downstream (mixed flow) flows. Variable AR is obtained by changing the (Dt) value in the designed ejector (Table 4). Convergence criteria of 10−6 are used for accuracy and good convergence. The mass flowrate variation with the iterations plot is used to find the minimal number of iterations necessary to reach convergence. The ER values are calculated by observing the changes in the m˙s and m˙p. The physical parameters such as molecular weight, viscosity, thermal conductivity, and specific heat of the refrigerants at the boundary conditions are taken from the refprop 10.0 database.

Table 3

Operating conditions

ParameterValue
Tcond (°C)32–47
Tevap (°C)0–14
ΔTevap (°C)10
Tgen (°C)69–90
ΔTgen (°C)20
AR3.52–5.33
NXP (mm)1–20
ParameterValue
Tcond (°C)32–47
Tevap (°C)0–14
ΔTevap (°C)10
Tgen (°C)69–90
ΔTgen (°C)20
AR3.52–5.33
NXP (mm)1–20
Table 4

Variable area ratio ejector specifications

Dt
(mm)
Dam
(mm)
AR = (Aam/At)
3.26.03.52
3.16.03.75
3.06.04.00
2.96.04.28
2.86.04.59
2.76.04.94
2.66.05.33
Dt
(mm)
Dam
(mm)
AR = (Aam/At)
3.26.03.52
3.16.03.75
3.06.04.00
2.96.04.28
2.86.04.59
2.76.04.94
2.66.05.33

2.4 Mesh Independence Test.

Increasing the number of cells yields more precise results, but it also lengthens the computational time. Meanwhile, the calculated results do not alter considerably beyond a particular grid density. Therefore, rigorous mesh independence tests are conducted to identify the appropriate grid size that takes less computational time and gives accurate results. A quadrilateral grid structure is adopted for constructing the grid (Fig. 2). The temperatures Tgen,Tevap, and Tcond are fixed to 90 °C, 5 °C, and 32 °C.
e(%)=|ER@103,950ERiER@103,950|×100
(9)

Equation (9) is used to calculate the relative error percentage in ERs for various grids mentioned in Table 5. Beyond grid number 22,335, the relative error in ER is less than 1%, and also, the axial static pressure distribution along the ejector length is comparatively close (Fig. 4). Hence, a quadrilateral mesh of size 22,335 is used for further investigations.

Fig. 4
Axial static pressure distribution with various grid numbers
Fig. 4
Axial static pressure distribution with various grid numbers
Close modal
Table 5

Relative error (%) between consecutive grid sizes

Grid sizeEntrainment ratioRelative error of ER
(%)
31500.25923.09
65500.28415.68
22,3350.33360.95
38,7500.33430.74
57,9650.33540.41
79,9800.33620.17
103,9500.3368NA
Grid sizeEntrainment ratioRelative error of ER
(%)
31500.25923.09
65500.28415.68
22,3350.33360.95
38,7500.33430.74
57,9650.33540.41
79,9800.33620.17
103,9500.3368NA

Note: Final grid size considered for the analysis is highlighted in bold.

2.5 Validation.

The current CFD model is validated using experimental findings of an R134a-based ejector [37]. Table 6 signifies the experimental working conditions adopted for verifying the numerical observations of the present model. ER variation with NXP is compared for two selective working conditions (Fig. 5). For both conditions, ER increases with an increase in NXP, showing good agreement with experimental findings. However, the disparity between the experimental and numerical outcomes is noticed to be within ±10% and ±5% for conditions 1 and 2, respectively. This minor deviation is mostly because of (a) the wall roughness of the experimental ejector and (b) the consideration of the working fluid as an ideal gas [45].

Fig. 5
Entrainment ratio comparison between simulation and experimental test results
Fig. 5
Entrainment ratio comparison between simulation and experimental test results
Close modal
Table 6

Experimental working conditions of the R134a ejector [37]

ConditionPgen
(bar)
Tgen
(°C)
Pevap
(bar)
Tevap
(°C)
Pcond(sat)
(bar)
Tcond(sat)
(°C)
122.13872.6556.5024
226.331004.25258.1032
ConditionPgen
(bar)
Tgen
(°C)
Pevap
(bar)
Tevap
(°C)
Pcond(sat)
(bar)
Tcond(sat)
(°C)
122.13872.6556.5024
226.331004.25258.1032

3 Results

In the ERS, the ejector is an essential component that regulates refrigerant flow across the evaporator and condenser. Therefore, a detailed performance study is conducted on the ejector for a wide variety of environment-friendly alternates of R134a refrigerant such as R1234yf, R1243zf, R152a, R513a, R440a, at a wide-ranging Tcond,Tevap, and Tgen. Furthermore, the study is extended to optimize the critical geometric parameters such as AR and NXP.

3.1 Effect of Operating Conditions.

The ejector geometry with an AR of four is adopted for investigating the influence of operating circumstances such as the Tcond,Tevap, and Tgen on the ER from Wang et al. [38].

3.1.1 Condenser Temperature Impact on Entrainment Ratio.

In line with topical climatic conditions, the Tcond is altered from 32 °C to 47 °C for investigating the critical point temperature corresponding to critical back pressure for all the selected refrigerants by keeping the Tgen and Tevap constant at 90 °C and 5 °C, respectively. For the refrigerants R1234yf, R1234zf, R152a, R513a, R134a, and R440a, the critical temperature is observed in the order of 36 °C, 36 °C, 38 °C, 38 °C, 42 °C, and 42 °C, respectively. The ER is maximum and constant for all the refrigerants below their respective critical temperatures. In this region, both primary and secondary flows are chocked, and they do not depend on Tcond. With the continuous increment of Tcond, only the primary flow is choked, whereas the secondary flow still depends on Tcond. Therefore, the ER is certainly decreased for all the refrigerants after their critical temperatures (Fig. 6). It is noticed that R1234yf and R1243zf give maximum ER for Tcond of 37 °C and followed by it drops. However for other refrigerants like R134a and R440a, a maximum ER is observed beyond Tcond of 40 °C. In between Tcond of 37 °C and 40 °C, the maximum ER is attained with refrigerants R152a and R513a.

Fig. 6
Change in entrainment ratio with condenser temperature
Fig. 6
Change in entrainment ratio with condenser temperature
Close modal

3.1.2 Impact of Evaporator Temperature on Entrainment Ratio.

Focusing on the automobile and industrial AC applications, the Tevap is varied from 0 °C to 14 °C to examine the ER at a fixed Tcond and Tgen of 32 °C and 90 °C, respectively. For a fixed Tgen, the primary flowrate, m˙p, remains constant. As the Tevap rises, the differential pressure between the secondary flow inlet and suction chamber increases, allowing more secondary flow, m˙s, into the ejector suction chamber. Therefore, the ER rises with the Tevap for all the refrigerants (Fig. 7). R1234yf and R1243zf refrigerants achieve the highest ER of 0.45 at 14 °C Tevap, while R134a has the lowest ER of 0.156 at 0 °C. Increasing Tevap from 0 °C to 14 °C raises the ER by approximately 50% for R1234yf, R1243zf, R152a, and R513a refrigerants while nearly doubling it for R134a and R440a.

Fig. 7
Change in entrainment ratio with evaporator temperature
Fig. 7
Change in entrainment ratio with evaporator temperature
Close modal

3.1.3 Generator Temperature Impact on Entrainment Ratio.

It is recommended to design the condenser in such a way that the Tcond should be above the ambient temperature in order to transfer heat through free convection. Also, the Tevap should not be too high or too low because it influences the power consumption. Therefore, a case study with the Tevap and Tcond of 5 °C and 32 °C, respectively, is conducted to identify the optimal Tgen for each refrigerant. At these operating conditions, the optimal Tgen for each refrigerant is identified to be in the range of 81–87 °C (Fig. 8). Optimal Tgen and corresponding maximum ER are identified for various refrigerants: R1234yf and R1243zf at 84 °C with 0.378 and 0.387, respectively, R440a at 84 °C with 0.344, R134a and R152a at 81 °C with 0.339 and 0.355, and R513a at 87 °C with 0.346.

Fig. 8
Change in entrainment ratio with generator temperature
Fig. 8
Change in entrainment ratio with generator temperature
Close modal

3.1.4 Pressure and Velocity Contours of All the Refrigerants.

Figure 9 represents the contours of pressure and velocity for all refrigerants at Tcond, Tevap, and Tgen of 32 °C, 5 °C, and 90 °C, respectively. For all, an array of expansion waves emerges from the primary nozzle exit and develops a converging duct to entrain the induced flow. Normal shock waves are formed at different regions for different refrigerants. In the case of R1234yf and R1243zf, the shock wave is almost disappeared, and the supersonic expansion wave is extended into the constant-area section exit with high velocity. The over-expansion of these expansion waves creates a flow field that pre-compresses the primary flow before mixing. The entrained induced flow combines with the primary flow and reduces its velocity. The combined flow experiences a weak normal shockwave in the diffuser and reaches the backpressure. This phenomenon can be the reason for the lesser critical back pressures of R1234yf and R1243zf than the other refrigerants, as shown in Fig. 6. For refrigerants R152a and R513a, an oblique shockwave is formed at the diffuser's inlet and propagates into the diffuser. In R134a and R440a, a convex-shaped shockwave is formed at the diffuser's inlet, and the velocity instantly falls to the subsonic levels.

Fig. 9
Pressure and velocity distribution across the ejector for different refrigerants
Fig. 9
Pressure and velocity distribution across the ejector for different refrigerants
Close modal

3.2 Effect of Geometrical Parameters.

The ejector's throat diameter,Dt, is altered from 3.2 mm to 2.6 mm (with 1 mm step reduction) to examine the influence of AR. The impact of the NXP on the ER is also investigated. In the following investigations, the Tcond, Tevap, and Tgen are fixed at 32 °C, 5 °C, and 90 °C, respectively, to run the ejector in the critical mode for all refrigerants.

3.2.1 Area Ratio Effect on Entrainment Ratio.

The effective flow area for the primary flow reduces with the reduction in throat diameter. As a result, more suction is created, and more induced flow is entrained. Thus the increase in AR increases the ER by decreasing the m˙p and increasing the m˙s (Fig. 10). For all ARs, R1234yf and R1243zf achieve higher and almost equal ER values. At an AR of 5.325, the maximum ER of 0.51 is attained with R1243zf and R1234yf. Conversely, R440a refrigerant yields the lowest ER of 0.13 at an AR of 3.516. As the AR increases from 3.52 to 5.325, around 82% increment in the ER is achieved with R1234yf, R1243zf, and R152A refrigerants, while it is raised 2.3 times with R134a, R513a, and R440a.

Fig. 10
Change in entrainment ratio with area ratio
Fig. 10
Change in entrainment ratio with area ratio
Close modal

3.2.2 Primary Nozzle Exit Position Effect on Ejector Performance.

At a given AR of four, as the NXP changes from 1 mm to 20 mm, the ER of all the refrigerants rises initially and is followed by reduction (Fig. 11). As the NXP moves away from the mixing section, the secondary flow's chocking position shifts toward upstream. Consequently, the hypothetical converging duct area of the secondary flow increases and allows more amount of induced flow. Therefore, the ER rises for all the refrigerants initially. If the NXP moves further, the primary flow's momentum decreases, resulting in a lower ER. It is not possible to define a single optimum NXP for all the refrigerants because of their difference in thermo-physical properties. For R1234yf, R1243zf, R513a, and R134a, the highest ER is achieved at 10 mm NXP. R440a achieves its peak at 15 mm NXP, while R152a's ER continuously increases within the tested NXP range. The optimum NXP for all the refrigerants is found to be in the range of 10–15 mm. At all the NXPs, the ER values are maximum with R1234yf refrigerant, and the highest ER of 0.355 is achieved at 10 mm NXP. Therefore, R1234yf is considered to optimize the NXP for a wide range of operating conditions.

Fig. 11
Change in entrainment ratio with nozzle exit position
Fig. 11
Change in entrainment ratio with nozzle exit position
Close modal

3.2.3 Optimum Nozzle Exit Position of R1234yf Refrigerant.

The effect of NXP on the ER for a wide variation of Tcond, Tevap, and Tgen is shown in Fig. 12. For all NXPs, the ER upsurges with a rise in Tevap and Tgen, in contrast, it decreases with an increase in Tcond. With the variation of NXP from 1 mm to 20 mm, the ER initially increases by 1.21 times and attains 0.35766 at NXP of 10 mm, and subsequently decreases by 0.91 times to the Tcond of 35 °C. However, for Tcond of 41 °C, a sharp increase of ER up to 7.06 times at NXP of 10 mm and decreases by 3.48 times to the peak value is observed. From this, it is obvious that the NXP influence on ejector performance is more significant at higher Tcond. A drop of 57.32% in the ER value can be noticed with a rise in Tcond from 35 °C to 41 °C at the optimum NXP of 10 mm.

Fig. 12
Change in entrainment ratio with nozzle exit position at various operating temperatures
Fig. 12
Change in entrainment ratio with nozzle exit position at various operating temperatures
Close modal

For a relatively higher Tevap of 6 °C and beyond, the ER decreases first and then increases to the peak value. However, at a lower Tevap of 0C, the ER increases initially, followed by decreases. The maximum ERs of 0.412, 0.372, and 0.330 for the Tevap of 9 °C, 6 °C, and 3 °C are achieved at 10 mm NXP. For Tevap of 0 °C and −3 °C, the peak ERs of 0.281 and 0.304 are obtained at 15 mm NXP. For every variation of Tevap around 3 °C, a minimum of 13% rise in ER is obtained at all NXPs.

The maximum ER at optimum NXP of 10 mm increases by 36.74% when the Tgen changes from 75 °C to 85 °C. At lower Tgen, the ER decreases significantly beyond the optimum value of NXP, which indicates that the impact of NXP is higher at lower Tgen. From this analysis, it can be comprehended that increasing NXP in a specific range can enhance the ER irrespective of the operating conditions. Moreover, the highest ER values are attained at 10 mm NXP. Hence, the optimum NXP for the present investigation is considered to be 10 mm.

3.2.4 Pressure and Velocity Contours of R1234yf Refrigerant at Various Nozzle Exit Positions.

The pressure and velocity contours and axial flow distribution throughout the ejector length with refrigerant R1234yf for various NXPs at Tgen, Tevap, and Tcond of 90 °C, 5 °C, and 32 °C, respectively, are described in Figs. 13 and 14. Diamond-shaped expansion waves of the primary jet are formed from the primary nozzle exit at all NXPs. In the case of 1 mm and 5 mm NXPs, the supersonic converging primary jet is extended to the CAM section exit with high velocity and replaces the normal shock wave. Consequently, the induced flow chokes while mixing with the primary flow and reduces its velocity. The combined stream decelerates to the subsonic speed in the diffuser section and reaches the backpressure. Additionally, for NXP of 5 mm, a minor shock wave is formed immediately after the CAM section exits. An increase in NXP reduces the over-expansion of the primary jet and develops a normal shockwave in the diffuser section. In the case of 10 mm, 15 mm, and 20 mm NXPs, the secondary flow gets choked at the mixing section entrance by the supersonic duct of the primary flow. For these, a convex-shaped normal shock wave is generated from the CAM section exit at a distance of 5 mm to 2 mm (Fig. 14). The changes in the mixing processes and the pressure differences of the two streams are responsible for the variation in ER with NXP. In addition, an increase in NXP than the optimum value results in an extensive loss of kinetic energy, which decreases the mass flowrate. It may be noted that the maximum suction-induced pressure differential poses the maximum ER at NXP of 10 mm.

Fig. 13
Pressure and velocity contours at various nozzle exit positions
Fig. 13
Pressure and velocity contours at various nozzle exit positions
Close modal
Fig. 14
Axial flow pressure and velocity distribution at different nozzle exit positions
Fig. 14
Axial flow pressure and velocity distribution at different nozzle exit positions
Close modal

3.2.5 Flow Characteristics of the R1234yf Refrigerant Along the Ejector Length.

Figures 15 and 16 illustrate the pressure, Mach number, and velocity contours of R1234yf refrigerant at Tcond,Tgen, and Tevap of 32 °C, 90 °C, and 5 °C, respectively. The high-pressure primary flow at 30.7 bar expands at the primary nozzle throat and attains the sonic velocity (Fig. 15). Due to the formation of oblique expansion shock waves in the primary jet core, the flow velocity at the primary nozzle exit is very high. This supersonic-primary jet acts as a wall and forms a convergent duct that creates suction of 1.75 bar in the mixing chamber to entrain the induced flow into the ejector (Fig. 15). The low-pressure (3.71 bar), low-velocity secondary flow is entrained through this effective area, gains momentum along with the primary flow and reaches a maximum velocity of around 637 m/s at a 5 mm distance from the entrance of constant-area section (Fig. 16). The over-expansion of this combined stream reduces its velocity and recompresses at the end portion of the pseudo shock region. At a 1.3 mm distance from the CAM section exit, a typical shock wave has formed and been transported into the diffuser, resulting in a sudden rise in pressure and a significant decrease in velocity. In the diffuser, all available kinetic energy is transformed into pressure energy, and the mixed flow reaches the condenser pressure of 8.23 bar with a velocity of 30.7 m/s.

Fig. 15
Pressure, Mach, and velocity contours of R1234yf ejector
Fig. 15
Pressure, Mach, and velocity contours of R1234yf ejector
Close modal
Fig. 16
Pressure and velocity variation of the primary and secondary flows along the ejector
Fig. 16
Pressure and velocity variation of the primary and secondary flows along the ejector
Close modal

4 Conclusions

The current study numerically evaluates the thermodynamic performance of an ejector for air-conditioning by comparing R134a with five zero-ODP and low-GWP alternatives, namely, R1234yf, R1243zf, R152a, R513a, and R440a under various operating conditions. The efficacy of the proposed model is validated against the experimental outcomes of a superheated ejector [37]. Besides this, the impact of design variables such as AR and NXP on the ER is also investigated using seven different ARs and five different NXPs for all refrigerants. The complex flow behavior inside the ejector is examined for all refrigerant types and NXPs, and optimal NXP is determined for various operating circumstances. Significant observations are as follows:

  • Refrigerants R1234yf and R1243zf yield the highest ER values in the critical mode of operation. However, the critical back pressure of these refrigerants is lower than that of others.

  • Increasing Tevap improves ER. All selected refrigerants outperform R134a in ER, with R1234yf and R1243zf achieving the highest values.

  • At the Tcond and Tevap of 32 °C and 5 °C, the optimal Tgen is identified in the range of 81–87 °C, corresponding to each refrigerant.

  • Increasing the AR from 3.52 to 5.325 increases the ER by about 82% with the refrigerants R1234yf, R1243zf, and R152a. However with refrigerants R134a, R513a, and R440a, the ER increases almost 2.3 times the initial value.

  • A single optimum NXP cannot be determined for all the refrigerants because of their difference in thermo-physical properties. For most refrigerants tested, the highest ER is achieved at 10 mm NXP. R440a achieves its peak at 15 mm NXP, while R152a's ER continuously increases within the tested NXP range.

  • At lower Tgen or higher Tcond, the impact of NXP is higher. Irrespective of the operating conditions for refrigerant R1234yf, the optimum NXP is obtained as 10 mm, which is 1.67 times the CAM chamber diameter.

From all the parametric investigations, among the tested refrigerants, R1234yf and R1243zf show excellent performance with the least ODP and GWP, and are proposed as suitable alternatives to R134a in ejector refrigeration systems for automotive or industrial applications.

Acknowledgment

The Department of Science and Technology (DST), New Delhi, under the Scheme for Young Scientists and Technologists (SYST) with DST Sanction Order No. SP/YO/565/2018, provides funding for the present work.

Conflict of Interest

There are no conflicts of interest.

Data Availability Statement

The datasets generated and supporting the findings of this article are obtainable from the corresponding author upon reasonable request.

Nomenclature

e =

relative error (%)

h =

enthalpy (J/kg)

u =

velocity (m/s)

A =

area of cross section (m2)

D =

diameter (m)

E =

overall energy (J)

L =

length (m)

P =

pressure (kPa)

T =

temperature (°C)

W =

work (kW)

=

mass flowrate (kg/s)

 Qe =

cooling capacity (kW)

Qg =

generator heat (kW)

Subscripts

d =

diffuser section

p =

primary flow

s =

secondary flow

t =

primary nozzle throat

am =

constant-area mixing section

ne =

primary nozzle exit

pm =

constant-pressure mixing section

gen =

generator

cond =

condenser

evap =

evaporator

Greek Symbols

αeff =

effective thermal conductivity (W/mK)

δij =

Kronecker delta function

Δ =

change in the value

μeff =

effective dynamic viscosity (Ns/m2)

ρ =

density (kg/m3)

τ =

stress tensor (Pa)

ω =

entrainment ratio

Acronyms

1D =

one-dimensional

3D =

three-dimensional

AR =

area ratio

CAM =

constant-area mixing

CFD =

computational fluid dynamics

COP =

coefficient of performance

CPM =

constant-pressure mixing

ER =

entrainment ratio

ERS =

ejector refrigeration system

GWP =

global warming potential

NXP =

nozzle exit position

ODP =

ozone depletion potential

VCRS =

vapor compression refrigeration system

References

1.
Lambert
,
M. A.
, and
Jones
,
B. J.
,
2006
, “
Automotive Adsorption Air Conditioner Powered by Exhaust Heat. Part 1: Conceptual and Embodiment Design
,”
Proc. Inst. Mech. Eng. Part D J. Automob. Eng
,
220
(
7
), pp.
959
972
.
2.
Talom
,
H. L.
, and
Beyene
,
A.
,
2009
, “
Heat Recovery From Automotive Engine
,”
Appl. Therm. Eng.
,
29
(
2–3
), pp.
439
444
.
3.
Tashtoush
,
B. M.
,
Al-Nimr
,
M. A.
, and
Khasawneh
,
M. A.
,
2019
, “
A Comprehensive Review of Ejector Design, Performance, and Applications
,”
Appl. Energy
,
240
, pp.
138
172
.
4.
Besagni
,
G.
,
Mereu
,
R.
, and
Inzoli
,
F.
,
2016
, “
Ejector Refrigeration: A Comprehensive Review
,”
Renew. Sustain. Energy Rev.
,
53
, pp.
373
407
.
5.
Wang
,
L.
,
Cai
,
W.
,
Zhao
,
H.
,
Lin
,
C.
, and
Yan
,
J.
,
2016
, “
Experimentation and Cycle Performance Prediction of Hybrid A/C System Using Automobile Exhaust Waste Heat
,”
Appl. Therm. Eng.
,
94
, pp.
314
323
.
6.
Ramesh
,
A. S.
, and
Joseph Sekhar
,
S.
,
2017
, “
Analytical and Numerical Studies of a Steam Ejector on the Effect of Nozzle Exit Position and Suction Chamber Angle to Fluid Flow and System Performance
,”
J. Appl. Fluid Mech.
,
10
(
1
), pp.
369
378
.
7.
Yapici
,
R.
,
Ersoy
,
H. K.
,
Aktoprakoǧlu
,
A.
,
Halkaci
,
H. S.
, and
Yiǧit
,
O.
,
2008
, “
Experimental Determination of the Optimum Performance of Ejector Refrigeration System Depending on Ejector Area Ratio
,”
Int. J. Refrig.
,
31
(
7
), pp.
1183
1189
.
8.
Dutton
,
J. C.
, and
Carroll
,
B. F.
,
1983
, “
Optimized Ejector-Diffuser Design Procedure for Natural Gas Vapor Recovery
,”
ASME J. Energy Resour. Technol.
,
105
(
3
), pp.
388
393
.
9.
Abbady
,
K.
,
Al-Mutawa
,
N.
, and
Almutairi
,
A.
,
2023
, “
The Performance Analysis of a Variable Geometry Ejector Utilizing CFD and Artificial Neural Network
,”
Energy Convers. Manage.
,
291
(
June
), p.
117318
.
10.
Seckin
,
C.
,
2017
, “
Parametric Analysis and Comparison of Ejector Expansion Refrigeration Cycles With Constant Area and Constant Pressure Ejectors
,”
ASME J. Energy Resour. Technol.
,
139
(
4
), p.
042006
.
11.
Aphornratana
,
S.
, and
Eames
,
I. W.
,
1997
, “
A Small Capacity Steam-Ejector Refrigerator: Experimental Investigation of a System Using Ejector With Movable Primary Nozzle
,”
Int. J. Refrig.
,
20
(
5
), pp.
352
358
.
12.
Hewedy
,
N. I. I.
,
Hamed
,
M. H.
,
Abou-Taleb
,
F. S.
, and
Ghonim
,
T. A.
,
2008
, “
Optimal Performance and Geometry of Supersonic Ejector
,”
ASME J. Fluids Eng.
,
130
(
4
), p.
0412041
.
13.
Li
,
C.
,
Li
,
Y.
,
Cai
,
W.
,
Hu
,
Y.
,
Chen
,
H.
, and
Yan
,
J.
,
2014
, “
Analysis on Performance Characteristics of Ejector With Variable Area-Ratio for Multi-evaporator Refrigeration System Based on Experimental Data
,”
Appl. Therm. Eng.
,
68
(
1–2
), pp.
125
132
.
14.
Huang
,
B. J.
,
Chang
,
J. M.
,
Wang
,
C. P.
, and
Petrenko
,
V. A.
,
1999
, “
1-D Analysis of Ejector Performance
,”
Int. J. Refrig.
,
22
(
5
), pp.
354
364
.
15.
Zhu
,
Y.
,
Cai
,
W.
,
Wen
,
C.
, and
Li
,
Y.
,
2009
, “
Numerical Investigation of Geometry Parameters for Design of High Performance Ejectors
,”
Appl. Therm. Eng.
,
29
(
5–6
), pp.
898
905
.
16.
Chen
,
W.
,
Shi
,
C.
,
Zhang
,
S.
,
Chen
,
H.
,
Chong
,
D.
, and
Yan
,
J.
,
2017
, “
Theoretical Analysis of Ejector Refrigeration System Performance Under Overall Modes
,”
Appl. Energy
,
185
(
Part 2
), pp.
2074
2084
.
17.
Chen
,
X.
,
Worall
,
M.
,
Omer
,
S.
,
Su
,
Y.
, and
Riffat
,
S.
,
2013
, “
Theoretical Studies of a Hybrid Ejector CO2 Compression Cooling System for Vehicles and Preliminary Experimental Investigations of an Ejector Cycle
,”
Appl. Energy
,
102
, pp.
931
942
.
18.
Bouhanguel
,
A.
,
Desevaux
,
P.
, and
Gavignet
,
E.
,
2011
, “
Flow Visualization in Supersonic Ejectors Using Laser Tomography Techniques
,”
Int. J. Refrig.
,
34
(
7
), pp.
1633
1640
.
19.
Little
,
A. B.
, and
Garimella
,
S.
,
2016
, “
Shadowgraph Visualization of Condensing R134a Flow Through Ejectors
,”
Int. J. Refrig.
,
68
, pp.
118
129
.
20.
Bartosiewicz
,
Y.
,
Aidoun
,
Z.
, and
Mercadier
,
Y.
,
2006
, “
Numerical Assessment of Ejector Operation for Refrigeration Applications Based on CFD
,”
Appl. Therm. Eng.
,
26
(
5–6
), pp.
604
612
.
21.
Bartosiewicz
,
Y.
,
Aidoun
,
Z.
,
Desevaux
,
P.
, and
Mercadier
,
Y.
,
2005
, “
Numerical and Experimental Investigations on Supersonic Ejectors
,”
Int. J. Heat Fluid Flow
,
26
(
1
), pp.
56
70
.
22.
Sriveerakul
,
T.
,
Aphornratana
,
S.
, and
Chunnanond
,
K.
,
2007
, “
Performance Prediction of Steam Ejector Using Computational Fluid Dynamics: Part 2. Flow Structure of a Steam Ejector Influenced by Operating Pressures and Geometries
,”
Int. J. Therm. Sci.
,
46
(
8
), pp.
823
833
.
23.
Rusly
,
E.
,
Aye
,
L.
,
Charters
,
W. W. S.
, and
Ooi
,
A.
,
2005
, “
CFD Analysis of Ejector in a Combined Ejector Cooling System
,”
Int. J. Refrig.
,
28
(
7
), pp.
1092
1101
.
24.
Pianthong
,
K.
,
Seehanam
,
W.
,
Behnia
,
M.
,
Sriveerakul
,
T.
, and
Aphornratana
,
S.
,
2007
, “
Investigation and Improvement of Ejector Refrigeration System Using Computational Fluid Dynamics Technique
,”
Energy Convers. Manage.
,
48
(
9
), pp.
2556
2564
.
25.
Wang
,
L.
,
Yan
,
J.
,
Wang
,
C.
, and
Li
,
X.
,
2017
, “
Numerical Study on Optimization of Ejector Primary Nozzle Geometries
,”
Int. J. Refrig.
,
76
, pp.
219
229
.
26.
UNEP
,
1987
, “
Montreal Protocol on Substances That Deplete the Ozone Layer
,” Final Act. United Nations.
27.
Bolaji
,
B. O.
,
2010
, “
Experimental Study of R152a and R32 to Replace R134a in a Domestic Refrigerator
,”
Energy
,
35
(
9
), pp.
3793
3798
.
28.
Seckin
,
C.
,
2020
, “
Effect of Operational Parameters on a Novel Combined Cycle of Ejector Refrigeration Cycle and Kalina Cycle
,”
ASME J. Energy Resour. Technol.
,
142
(
1
), p.
012001
.
29.
Sánchez
,
D.
,
Cabello
,
R.
,
Llopis
,
R.
,
Arauzo
,
I.
,
Catalán-Gil
,
J.
, and
Torrella
,
E.
,
2017
, “
Energy Performance Evaluation of R1234yf, R1234ze(E), R600a, R290 and R152a as Low-GWP R134a Alternatives
,”
Int. J. Refrig.
,
74
(
2017
), pp.
267
280
.
30.
Khaliq
,
A.
,
Almohammadi
,
B. A.
,
Alharthi
,
M. A.
,
Siddiqui
,
M. A.
, and
Kumar
,
R.
,
2021
, “
Investigation of a Combined Refrigeration and Air Conditioning System Based on Two-Phase Ejector Driven by Exhaust Gases of Natural Gas Fueled Homogeneous Charge Compression Ignition Engine
,”
ASME J. Energy Resour. Technol.
,
143
(
12
), p.
120911
.
31.
Fang
,
Y.
,
Croquer
,
S.
,
Poncet
,
S.
,
Aidoun
,
Z.
, and
Bartosiewicz
,
Y.
,
2017
, “
Drop-In Replacement in a R134 Ejector Refrigeration Cycle by HFO Refrigerants
,”
Int. J. Refrig.
,
77
, pp.
87
98
.
32.
Modi
,
N.
,
Pandya
,
B.
,
Patel
,
J.
, and
Mudgal
,
A.
,
2019
, “
Advanced Exergetic Assessment of a Vapor Compression Cycle With Alternative Refrigerants
,”
ASME J. Energy Resour. Technol.
,
141
(
9
), p.
092002
.
33.
Mastrowski
,
M.
,
Smolka
,
J.
,
Butrymowicz
,
D.
,
Antczak
,
L.
,
Gagan
,
J.
, and
Śmierciew
,
K.
,
2024
, “
Experimental Validation of the Theoretical Ejector Model in a Low-Grade Waste Heat Refrigeration System Using R1233zdE as a Working Fluid
,”
Appl. Therm. Eng.
,
236
(
PD
), p.
121716
.
34.
İşkan, Ü.
, and
Direk
,
M.
,
2022
, “
Experimental Performance Evaluation of the Dual-Evaporator Ejector Refrigeration System Using Environmentally Friendly Refrigerants of R1234ze(E), ND, R515a, R456a, and R516a as a Replacement for R134a
,”
J. Clean. Prod.
,
352
(
Jan.
), p.
131612
.
35.
Seckin
,
C.
,
2024
, “
Investigation on the Effects of Nanorefrigerants in a Combined Cycle of Ejector Refrigeration Cycle and Kalina Cycle
,”
ASME J. Energy Resour. Technol.
,
146
(
2
), p.
012401
.
36.
Colombo
,
L. P. M.
,
Lucchini
,
A.
, and
Molinaroli
,
L.
,
2020
, “
Experimental Analysis of the Use of R1234yf and R1234ze(E) as Drop-In Alternatives of R134a in a Water-to-Water Heat Pump
,”
Int. J. Refrig.
,
115
, pp.
18
27
.
37.
Falsafioon
,
M.
,
Aidoun
,
Z.
, and
Poirier
,
M.
,
2017
, “
A Numerical and Experimental Study of Ejector Internal Flow Structure and Geometry Modification for Maximized Performance
,”
IOP Conf. Ser. Mater. Sci. Eng.
,
280
(
1
), p.
012011
.
38.
Wang
,
H.
,
Cai
,
W.
,
Wang
,
Y.
,
Yan
,
J.
, and
Wang
,
L.
,
2016
, “
Experimental Study of the Behavior of a Hybrid Ejector-Based Air-Conditioning System With R134a
,”
Energy Convers. Manage.
,
112
, pp.
31
40
.
39.
Sankarlal
,
T.
, and
Mani
,
A.
,
2007
, “
Experimental Investigations on Ejector Refrigeration System With Ammonia
,”
Renew. Energy
,
32
(
8
), pp.
1403
1413
.
40.
Samaké
,
O.
,
Galanis
,
N.
, and
Sorin
,
M.
,
2016
, “
On the Design and Corresponding Performance of Steam Jet Ejectors
,”
Desalination
,
381
, pp.
15
25
.
41.
Ramesh
,
A. S.
, and
Sekhar
,
S. J.
,
2018
, “
Experimental and Numerical Investigations on the Effect of Suction Chamber Angle and Nozzle Exit Position of a Steam-Jet Ejector
,”
Energy
,
164
, pp.
1097
1113
.
42.
Ji
,
M. K.
,
Utomo
,
T.
,
Woo
,
J. S.
,
Lee
,
Y. H.
,
Jeong
,
H. M.
, and
Chung
,
H. S.
,
2010
, “
CFD Investigation on the Flow Structure Inside Thermo Vapor Compressor
,”
Energy
,
35
(
6
), pp.
2694
2702
.
43.
Varga
,
S.
,
Oliveira
,
A. C.
, and
Diaconu
,
B.
,
2009
, “
Influence of Geometrical Factors on Steam Ejector Performance – A Numerical Assessment
,”
Int. J. Refrig.
,
32
(
7
), pp.
1694
1701
.
44.
Chong
,
D.
,
Hu
,
M.
,
Chen
,
W.
,
Wang
,
J.
,
Liu
,
J.
, and
Yan
,
J.
,
2014
, “
Experimental and Numerical Analysis of Supersonic Air Ejector
,”
Appl. Energy
,
130
, pp.
679
684
.
45.
Mazzelli
,
F.
, and
Milazzo
,
A.
,
2015
, “
Performance Analysis of a Supersonic Ejector Cycle Working With R245fa
,”
Int. J. Refrig.
,
49
, pp.
79
92
.