I. INTRODUCTION

THE boiling process using nanoparticles is one of the most critical heat transfer mechanisms, which plays a significant role in industrial sectors, such as cooling systems, power plants, and chemical reactors ^{(1)}. The industrial sector continually seeks new ways to improve the heat transfer properties of working fluids in refrigeration systems.

Nanofluid boiling is considered an important research study that offers various opportunities to explore new frontiers. However, it also comes with great challenges. For several years, some studies have been based on nanofluids for the improvement of heat transfer during the boiling process, however, the data on the boiling heat transfer coefficient and the critical heat flux have been unpredictable ^{(2)}, ^{(3)}, ^{(4)}.

Wang et al. ^{(5)} point out that the usefulness of nanofluids in refrigeration systems is considered a potential way to optimize the energy efficiency and reliability of heating, ventilation, air conditioning and refrigeration installations (HVAC & R) and to make the use of environmentally friendly refrigerants economical.

Yin et al. ^{(6)} established a nanofluid model consisting of suspended nanoparticles for the improvement of heat transfer of boiling flow. The results reveal that both flow rate and heating temperature have an effect on the heat transfer from boiling flow to lower heating temperature. On the other hand, they obtained that, with an increase in the heating temperature, the suspended nanoparticles have much more abrupt rotational and translational movements, which can significantly improve the heat transfer of the nanofluids.

Mukthiyar et al. ^{(7)} point out that the most used refrigerant is the R410A refrigerant. This is a mixture of difluoromethane (CH2F2) and pentafluoro methane (CHF2CF3), being successfully marketed in air conditioning. On the other hand, Heredia et al. ^{(8)} mention that the refrigerant R410A has a high global warming potential (GWP) of 2 088. However, Fannon et al. ^{(9)} suggest that R410A refrigerant has better performance and minimizes pressure drop, especially for high refrigerant flow rates. Which indicates that R410A would be a better option for the design of new cooling systems direct expansion (OX).

Shao et al. ^{(10)} carried out a numerical study on the heat transfer of R410A during the boiling flow. In this study thermal phase change model in fluent was used. The results of this research show that the heat transfer coefficient increased with the mass flow growth, while it decreased with the growth of vapor quality due to the interaction of nucleate boiling and fluid convection. However, the effect of heat flux was slight on the heat transfer coefficient of boiling flux, indicating that boiling flux is mainly governed by fluid convection.

On the other hand, Nair et al. ^{(11)} in their study of nanorefrigerants: A Comprehensive Review of its Past, Present, and Future, point out that nanorefrigerants can greatly reduce the energy consumption of a cooling system. The show that improved thermal conductivity of a nano refrigerant is only partially responsible for a higher boiling heat transfer coefficient, also show the COP of the refrigeration system increases with the addition of nanoparticles in the refrigerant. Finally, the authors mention that for the R410A refrigerant nano lubricants based on polyol oils should be used ester (POE) due to its miscibility.

Low global warming potential (GWP) refrigerants, such as R600a, have become a very important topic of study. That because they significantly reduce the emission of greenhouse gases from industries dedicated to air conditioning ^{(12)}. One example of that is the research developed by Longo et al. ^{(13)} carried out the thermodynamic and heat transfer evaluation of low-GWP refrigerants such as R600a, R1234ze (Z), and R1233zd (E) as an alternative to traditional low-pressure HFC refrigerants, such as R245fa, for heat pump (HP) and organic Rankine cycle (ORC) applications.

Gobinath and Venugopal ^{(14)} claim that CuO nanoparticles of 0.1 % volume fraction in R600a established nearly 20 % enhancement in heat transfer coefficient of the refrigerant at higher heat flux. The optimum concentration of CuO nanoparticles in stable dispersion for a prolonged time and contribute for significant enhancement in the heat transfer coefficient of R600a refrigerant is experimentally found as 0.05 % by volume.

The incorporation of solid nanoparticles into common fluids results in the formation of a nanofluid which can be considered an interesting technique to improve the thermal characteristics of a working fluid ^{(15)}. Akhavan-Behabadi et al. ^{(16)} executed an experimental study on the effect of CuO nanoparticles on the boiling flux of the mixture (R600a/oil) inside a smooth horizontal tube. These experiments were carried out under parameters of mass velocities of 50 to 400 kg/m2s, inlet vapor qualities of 0 to 0.9, heat fluxes of 3 to 8 kW/m2 and mass fractions of CuO nanoparticles from 0 to 1.5 % by weight. The mixing of the nanoparticles with the base fluid was carried out by passing them through a system at high speed for approximately 3 hours. The results reveal that the addition of CuO nanoparticles significantly improves heat transfer by up to 63 % relative to the heat transfer coefficient of (R600a/oil) without nanoparticles.

In the research carried out by Rabiee and Atf ^{(17)}, a numerical study was carried out using computational fluid dynamics (CFO), in which AZ2O3 and CuO nanoparticles were incorporated into the fluid, basis to achieve an increase in the heat transfer coefficient during the boiling process. For this study, the Reynolds-Averaged Navier-Stokes equations were used accompanied by a mechanistic model developed by the Rensselaer Polytechnic Institute (RPI) to simulate the boiling flow field with an Eulerian-Eulerian approach for each phase. They concluded that, copper oxide compared to alumina nanoparticles would lead to higher amounts of heat transfer coefficients along pipes.

Sharif et al. ^{(18)} developed a mechanism to improve the performance of the refrigeration system through the use of nanorefrigerants and nanolubricants. This study mention that, the use of nanorefrigerants and nanolubricants in the vapor compression refrigeration system (VCRS, for its acronym in English) increased heat transfer coefficients from 12 % to 101 % and improved thermal conductivity by up to 4 %. The solubility and miscibility of the coolant-oil mixture with nanoparticle additives was improved by up to 12 % and showed a 24 % COP improvement. On the other hand, Senthilkumar ^{(19)} in their prospective study of nanolubricants and nanorefrigerants on energy savings in the vapor compression refrigeration system point out that, in the R410A refrigeration system, COPs of 4 were obtained, likewise a 8 % with 0.1 % and 0.5 % and 7 % cooling capacity using diamond nanoparticles.

Flow patterns are considered to be the configurations of a fluid that are formed when transporting two or more phases together through a pipe. These have many applications in industry, such as nuclear power plants, heat exchangers and chemical reactors ^{(20)}, ^{(21)}.

Lee et al. ^{(22)} conducted an investigation on the boiling flow patterns and drying characteristics of R-1234ze (E) in a plate exchanger (PHE), classifying these patterns into three flow regimes which are: rough liquid film flow, pulsating annular and stable annular flow. The boiling flow pattern map of R- 1234ze (E) in the exchanger was carried out in terms of liquid and vapor surface moments. Transitions of flow regimes on the map were expressed in terms of dimensionless numbers. Furthermore, based on the relationship between heat transfer coefficient and flow patterns in the PHE, stable annular flow was recommended as a preferable pattern to achieve very high heat transfer coefficients.

Yang et al. ^{(23)} carried out an experimental study of the flow patterns of R600a in a smooth horizontal tube with an internal diameter of 6 mm using a high-speed camera, where four main flow regimes could be observed: piston flow, corrugated- stratified, slug and annular. The experiments were carried out under conditions of saturation pressure of 0.215 to 0.415 MPa, mass fluxes of 67 to 194 kg/m2s and heat fluxes of 10.6 to 75.0 kW/m2, respectively. Eight correlations were evaluated showing that the Liu and Winterton correlation provides the best fit to the experimental data with a mean absolute relative deviation of 11.5 %.

In another investigation, Copetti et al. ^{(24)} carried out an experimental study of the boiling flow of the refrigerant R600a in an aluminum tube of 1.47 mm hydraulic diameter. This study was carried out considering heat fluxes in the range of 5 to 30 kW/m2, mass velocities adjusted to discrete values in the range of 50 to 200 kg/m2s and a saturation temperature of 20 °C. In addition, they presented some images of flow patterns, in different conditions, the main patterns identified were slug, intermittent and annular. The results show that the heat coefficient increases with increasing heat flux, and when the mass velocity (G) is high the growth effect is even greater.

Abdollahi et al. ^{(25)} carried out a numerical investigation to study heat transfer characteristics of nanofluids flowing through heat sink having a V-type inlet and outlet arrangement. SiO-2 - water, AI2O3 - water, ZnO - water, and CuO - water nanofluids were used as working fluids in the investigation. The average diameter size of the nanoparticles used was 30 nm, 40 nm, and 60 nm and volumetric concentration varied from 1 % to 2 %.

The objective of this research is to know and understand how the flow patterns of two highly used refrigerants are affected when they are mixed with nanoparticles.

II. METHODS AND METHODOLOGY

Nanorefrigerants are mainly used in low temperature applications, such as air conditioning, refrigeration systems and vapor compression systems ^{(26)}. The addition of nanoparticles, especially metal oxides in the refrigerant, varies the properties and increases the heat transfer performance, improving the heat transfer performance of refrigeration systems ^{(27)}.

*A. Thermophysical Properties*

Thermophysical properties represent important parameters that are obtained using the Engineering Equation Solver (EES) software. Therefore, the corresponding properties of the R600a refrigerant in liquid and vapor states are observed in Table I.

The thermodynamic properties of R410A as a zeotropic refrigerant are shown in Table II, for both the liquid and vapor phases. Table III.

*B. Analytical Model*

To obtain the flow patterns, dimensionless variables proposed in the study carried out by Taitel-Dukler are used, which have been modified in the research by Kattan et al. ^{(30)}, therefore it is necessary to define equations 1, 2, 3, 4, 5 and 6:

Where, hL is the height of the liquid (m), D is the internal diameter of the tube (m), PL is the wet part of the perimeter(m), while, Pv is the complementary part of the perimeter in contact with the vapor (m), AL and Av are the corresponding cross-sectional areas (m2) and Pi is the phase interface (m).

On the other hand, for the calculation of the average local void fraction from the steam quality, derived from the work carried out by Rouhani and Axelsson^{(31)}for horizontal pipes in a section void fraction transversal, e*.*Equation 7.

Where, x is the vapor quality, Pv and PL are the density of the vapor and liquid respectively (kg/m3), g is the acceleration of gravity (m/s2), σ is the surface tension (N/m) and G is the total mass velocity of liquid and vapor (kg/sm2).

The average local void fraction model provides an explicit function of the total mass flux. Therefore, the surface section of tube A, the values of*A*
_{
LD
} and*A*
_{
VD
} are directly identifiable. Equations 8 and 9.

On the other hand, the dimensions of the liquid height hLD and the dimensionless length of the liquid-vapor interface PiD can be expressed as a function of the stratified flow angle θ strat· Where, qstrat is the stratified flow angle of the tube perimeter (rad), it can be determined from an approximate expression, evaluated by Biberg ^{(32)}. Equation 10.

Consequently, since the void fraction e is a function of the mass velocity G, it influences the position of the transition curves in the flow maps, proposed by Klimenko-Fyodorov ^{(33)}. Equation 11.

Where, bLa is the Laplace constant.

*C. Stratified and Intermittent/annular Flow Pattern*

This flow pattern occurs at very low G mass velocities when the Kelvin-Helmholtz instability criterion is counteracted by viscous forces. It is done using the equation 12 proposed by Kattan et al. ^{(30)}.

Where, Gstrat is the stratified flow transition mass velocity (kg/m2s).

The annular flow pattern is obtained when the liquid wets the entire periphery of the tube with the vapor flowing in the center of the tube ^{(34)}. The annular flow pattern has been considered to be achieved when the movement of the liquid flowing at the top of the tube ^{(35)}.

The intermittent flow pattern occurs at low temperature, therefore, between intermittent and annular flow is a function of the void fraction ^{(35)}. It is defined by a fixed Martinelli parameter Xtt = 0.34

*D. Stratified-Wavy and Drying-Fog Flow Pattern*

It is characterized by a wavy interface of the liquid, where waves exist and these are of reduced magnitude and cannot reach the upper part of the tube ^{(35)}.

It is calculated from the original expression of Kattan et al. ^{(30)}. Equation 13.

Where, Gwavy is the wave flow transition mass velocity kg/m2s. However, to obtain the Froude number the following expression is used. Equations 14 and 15:

Where, FrL and FrV is the Froude number for liquid and vapor, respectively.

Kattan et al. ^{(36)} used the following equations to find the transition limits of drying and fog flux. Equations 16 and 17.

Where, Gdryout is the drying transition mass velocity (kg/m2s), Gmist is the fog flow transition mass velocity (kg/m2s) and q is the fog flux local heat (W/m2). The approach of Mori et al. ^{(37)} includes the new drying limits in his research and they can be calculated from the equations 18 and 19:

Therefore, to calculate the critical heat flux qcrit (kg/m2s) the expression, given by Kutateladze ^{(38)} is used. Equations 20.

*E. Numerical Model*

ANSYS CFD is a tool that offers higher accuracy quantitative forecasts of fluid interactions and connections ^{(39)}.

The pipes are manufactured to transport fluids inside, so for this study a square horizontal pipe is used, whose design parameters are shown in Table IV.

For this purpose, within the Fluent setup, the materials and edge conditions are configured for correct operation. Below, Table V shows the simulation parameters.

The volume fractions for both liquid and vapor phases are defined by the equation 21:

Where, the volume divisions refer to the space involved in each phase, and the laws of conservation of mass and moment are fulfilled in each stage exclusively ^{(39)}.

Equation of conservation of mass or also called the equation of continuity, this defines the increase and decrease of mass that occurs in the phase change (liquid-vapor) taking into account the principle of conservation of mass. Equation 22.

Equation 23 for this balance shows the external forces acting within a phase change, which correspond to the movement of the mass, which are produced due to fluid turbulence. Equation 23.

*F. Mesh Generation*

The geometry mesh is implemented for assigning the boundary conditions, using a body size of 1.5 mm, as can be seen in Figure 1.

Obtaining its quality using the Skewnees tool within Fluent’s metric meshing, the same as seen in Figure 2, this type of quality states that as long as the number of elements and nodes is located in a range of 0 to 0.5, the meshing will converge from successful way.

III. RESULTS AND DISCUSSION

The flow maps of R600a refrigerant are compared with the research of Yang et al. ^{(23)} who studies pure flow maps. The flow maps for R600a/CuO upon adding 3 % and 5 % nanoparticles in concentration as shown in Figure 3 a) and b), the dimensions of diameter, mass velocity and heat flux are kept constant. Furthermore, it can be seen that the slug/stratified-wavy flow regime zone is more extensive, since the transition fine between slug and slug+SW has a higher mass velocity.

To check the accuracy of the flowchart, it is compared with the work of Hu et al. ^{(40)} who carried out a study of the flow diagrams of pure R410A refrigerant and R410A containing lubricating oil. In Figure 4 a) and b), the comparison of the flow maps for pure R410A and R410A with 3 % and 5 % nanoparticles, in which it is possible to see that the vapor quality is lower at the flow interface ( I/A), however, this is directed to the left, thus accelerating the phase change process, because, when adding nanoparticles, the quality varies by 10 %, and reaching the drying regime is faster than for the pure refrigerant, in the same way as for the fog zone.

Figure 5 indicates that the flow maps for R600a/CuO with 3 % nanoparticles, the transition line between slug and slug+SW has a higher mass velocity. While, by adding 5 % of nanoparticles, a greater heat transfer occurs, causing an intersection between the zones of the annular intermittent flow.

Figure 6 shows the simulation results for the R410A refrigerant with a mixture of CuO nanoparticles, where the heat transfer is dominated by subcooled boiling and nucleated boiling, which is why the phase change when adding a 3 % and 5 % CuO nanoparticles are produced at lower vapor qualities, thus generating the transition limit (I/A) to develop early in the boiling process.

IV. CONCLUSIONS

The following conclusions are drawn by studying the flow patterns in R600a and R410A refrigerants in pure state and with nanoparticles:

Through an analytical study, it was proven that by adding nanoparticles in concentrations of 3 % to 5 % to the R600a and R410A refrigerants, the transition between the flow regimes increases by 25 %, while the vapor quality is reduced. For this reason it is deduced that its heat transfer increases and the phase change process begins to occur at short lengths within the pipe.

In the numerical study carried out it was concluded that in the R600a refrigerant the flow boiling heat transfer coefficient increased from 4500 to 5400*kg/m*
^{2}
*s*using CuO nanoparticles. Also the heat transfer coefficient was improved in regimes such as slug and stratified-wavy (SW). That occur at vapor qualities between 0.1 to 0.5 and as its quality increased this coefficient decreased from 3600 to 2400*kg/m*
^{2}
*s.*The result is drying (D) and fog (M) regimes occupying a vapor volume of less than 20 %.

Through an analytical and numerical study, the flow boiling heat transfer performance of CuO nanoparticles in the refrigerants R600a and R410A was analyzed.