Fu Yijun
College of Environment and Building, University of Shanghai for Science and Technology, Shanghai, PR. China
Li Guo
College of Environment and Building, University of Shanghai for Science and Technology, Shanghai, PR. China
Lv Jing
College of Environment and Building, University of Shanghai for Science and Technology, Shanghai, PR. China

Keywords: CO2, Heat transfer characteristics, Microchannel evaporator, Simulation model



Simulating model based on a specific CO2 microchannel evaporator was established through control volume method with MATLAB, in which both wet and dry conditions for air side, and two-phase and over-heat zones for CO2side have been considered during the evaporative process. Simulation results showed little discrepancy with pervious experimental data which validates the model. And then the heat transfer characteristics in microchannel evaporator were simulated under different inlet air parameters. It was shown that air velocity has the greatest impact on heat transfer effect, followed by air temperature, and air humidity at last. Meanwhile, the dry-out point also has an important impact on heat transfer performance: before the dry out happens, the heat transfer coefficient of the CO2 increased with higher air temperature, relative humidity and velocity, while after the dry out occurs, there has been a drastic decline of convective heat transfer coefficient. Therefore, the dry-out point should be postponed for better performance. Then, structural optimization has been made by utilizing two-stage series evaporators. Corresponding simulation results showed that 37.5% area of the original experimental device can still achieve 90.5% heat transfer rate of the former one. So, this method can greatly improve the heat transfer effect of the CO2 microchannel evaporator.


The main devices of heat transfer in the refrigerating cycle of CO2 have been going through the development from the finned tube style to the microchannel style. Compared to the traditional heat exchanger, the microchannel heat exchanger is usually smaller and higher heat transfer coefficient, but it’s pressure resistance and drop is higher, which may easily cause congestion and imbalanced distribution of fluid. CO2 can cover the shortage of microchannel heat exchanger due to its low ratio between liquid and gas density [1]. However, when the hydraulic diameter is smaller than 3mm, the two-phase flow and heat transfer regulation differs from the normal size. More noticeable microscale effect can be observed in narrow passageway [2].

Many research institutions have studied on this issue that mainly focus on the boiling heat transfer coefficient of the two-phase field, critical heat flux, dry-out point, two-phase flow pattern and pressure drop model [3]. Cheng et al. have discovered that the critical dryness of carbon dioxide was generally between 0.5 and 0.7, which was much lower than that of R22 with a critical dryness usually between 0.8 and 0.9 [4]. Then, they have considered the characteristics of intermittent flow, annular flow, dry-out inception and mist flow to modify the boiling heat transfer correlation under the basis research of Wojtan [5]. Zhang has established a two-dimensional distributed parameter model for the CO2 microchannel evaporator and proposed a modified heat transfer correlation after comparing to the experimental data [6].

Several appropriate heat transfer correlations are selected according to the heat transfer characteristics of CO2 in microchannel evaporator, and comprehensively considered the different heat transfer characteristics of wet and dry conditions on air side along with two-phase region and overheated region of CO2. Parameter distribution simulation model of CO2 microchannel evaporator has been established and verified through the experimental results. Finally, structural optimization has been proposed and verified through further simulation under different channel number, air temperature, humidity and velocity have been analyzed for studying their impact on heat transfer performance.

I. Experiment Research

A. Microchannel evaporator

Experiment research on a parallel flow micro-channel evaporator which is composed of 36 parallel flat tubes, each of which has 18 microchannels with equivalent diameter of 1.096 mm. The two-phase CO2 coming from the collecting pipes flows into the microchannel and exchanges heat with the air in the louver fin between the microchannels. Figure 1is the structure of the microchannel evaporator and it’s detailed 3D diagram. The calculated main structural parameters are shown in Table 1.

Figure 1. Diagram of the microchannel evaporator.


Table 1. Main structural parameters of microchannel evaporator.

Upwind surface

Air direction depth

Volume (cm³)

Heat exchange area (m²)

Equivalent diameter

Air side

Refrigerant side








B. Experimental system

The experimental table of the CO2 microchannel evaporator was set up (see Figure 2). The conditions of the evaporator side are provided by the Enthalpy Different Laboratory. The platinum resistances and pressure transmitters were installed in the evaporator inlet and outlet in order to measure the temperature and pressure of CO2. Thermocouples were fixed on the surface of the evaporator to measure it’s tube temperature. The temperature, humidity and speed of air side were measured by thermometer, hydrometer and anemometer. Finally via the electronic expansion valve, the dryness and mass flow rate of CO2 at the inlet of the evaporator was adjusted.

Figure 2. The experiment system diagram.


C. Experimental results

The 18th flat tube was analyzed and divided into 9 sections, which is 90 mm with measuring points set in the center of each section. The incipient air temperature is set to 23°C, and relative humidity is 25%, so the dew point temperature is 2.14°C.The experiment measured CO2 mass flow rate, inlet dryness, pressure, evaporation temperature, wall temperature, air temperature, humidity and speed which are shown in Table 2 and Table 3. The convective heat transfer coefficient and heat transfer amount of each section of this flat tube would be calculated according to the experimental data, which are shown in Table 2 and Table 3. Table 4 is the convective heat transfer coefficient and heat transfer amount along the length distribution.

Table 2. Experimental measurement values of CO2 side.


Mass flow

Inlet dryness

Inlet pressure

Outlet pressure

Outlet Temperature

Measured values







Table 3. Distribution of Parameters.

Measuring points









Wind speed (m/s)









Outlet temperature (°C)









Outlet humidity (%)









Wall temperature (°C)










Table 4. The convective heat transfer coefficient and heat transfer amount along the length distribution.

Length (mm)










Convective heat Transfer coefficient (W/m² K)










Heat exchange amount (W)











II. Simulation Model

A. Heat transfer correlation of CO2 side

1)    Overheated region

According to the different evaporator outlet states of CO2, the refrigerant flow can be divided into two-phase region and overheated region. As for the overheated region, different heat transfer correlations were selected according to the Reynolds number: When Re ≥ 2300, the convective heat transfer coefficient was calculated by Gnielinski formula [7]; When Re < 2300, the convective heat transfer coefficient was calculated by Sieder-Tate formula [8].

2)    Two-phase region

The currently available CO2 boiling heat transfer correlations are mainly Shah, Gungor and Winterton, Hwang, Yoon, and Cheng correlations. In Cheng correlation, the whole two-phase region is divided into 3 phases as the intermittent flow, annular flow and mist flow according to the boundary point of intermittent flow and annular flow and the dry-out point. Compared with the experimental data in the reference [9], the Cheng correlation was considered to be the most accurate in this experimental condition. So the Cheng correlation was selected in our simulation.

B. Heat transfer correlation of air side

When the wall temperature is below the air dew point temperature, dew will occur on the surface of the flat tube. So the dry and the wet working conditions should be both considered to analyze the heat transfer on the air side. Many research focus on dry condition, while regardless of wet condition. In the wet condition, the surface thermal resistance increases and the heat transfer coefficient will be much smaller. The correlations developed by Kim and Bullard can predict accurately about the heat and mass transfer performance of the shutters both in dry and wet conditions [10]. So their correlation was used in our simulation model.

C. Controlling equations

In order to make the process of calculation easier, the mathematical model of CO2 microchannel evaporator was assumed as follows:

1)    CO2 is equally divided into flow each microchannel;

2)    No thermal conduction or heat resistance exists between microchannels;

3)    CO2 side and air side are both steady flow;

4)    The air on the condensation water surface is saturated and the thermal resistance of the condensed water is negligible;

5)    The effect of lubricating oil and noncondensing gas is not considered.

Making the flat tube and the 1/2 louver fin on the upside and underside of it as the research object, the control unit is shown in Figure 3.

Figure 3. Sectional view of control unit.

As shown in Figure 4, every control volume can be regarded as a small cross flow heat exchanger, which was analyzed by energy conservation:

Air side heat exchange:


CO2 side heat exchange (two-phase):


CO2 side heat exchange (overheated):


Where Q represents heat transfer rate (W), M the mass flow rate, (kg/s), h the enthalpy(kJ/kg), x thedryness,and I the latent heat of vaporization(kJ/kg).

Where subscript abbreviation a represent air, r refrigerant, i inlet, o outlet, tp two-phase, j the j control volume.

Figure 4. The figure of control volume.


Heat transfer between air and pipe wall under dry and wet conditions:





Where α represent sensible heat transfer coefficient (W/m²·k), β the mass transfer coefficient (kg/m²·s), η the cooling efficiency, A the heat transfer area (m²), and T the temperature(K).

Where subscript abbreviation: d represents dry air, w the water film, m average, and s saturated.

D. Simulation process design

The heat transfer existing both in two-phase region and overheated region in normal heat pump system, because a certain overheated degree at evaporator outlet is usually required in order to ensure CO2 enter into the compressor with gas phase. In this condition, the point with dryness equal 1 of CO2 was calculated first to divide the heat transfer process into the two-phase region and the overheated region. The specific calculation process is shown in Figure 5, in which the state parameters of CO2 and air were obtained from MATLAB by manipulating the REFPROP.

Figure 5. Simulation process of heat transfer.

II. Results and Discussion

A. Comparison of experimental and simulation results

Comparison between simulating and experimental values of CO2 temperature, tube wall temperature, air inlet and outlet temperature and the convective heat transfer coefficient are respectively shown in Figure 6 and Figure 7.

Figure 6. Comparison of CO2 temperature, wall temperature, air inlet and outlet temperatures.

Figure 7. Comparison of the convection heat transfer coefficient.

The relative errors between experimental and simulated values were calculated. The relative error of CO2 temperature and wall temperature is under 10%.But the relative error of convective heat transfer coefficient is about 18%. This is because that microchannel evaporator tended to have an uneven flow distribution problem during actual operation which was omitted in the model. Which may cause too much or less CO2 mass flow rate in some microchannes. Meanwhile, similar changing trend of simulated and experimental values can be seen from Figure 7 with acceptable errors range, which is usually 20% for a simulation of engineering application. So the simulation model is highly reliable forfurther analysis and optimization.

B. Sturctural optimization

Considering that the location of the drying point is at the first 1/4 length in the channel as is shown in Figure 7, heat transfer efficiency decreases sharply at the latter part of the channel. Therefore, we divide the former one evaporator into two ones with a gas-liquid separator installed between them. The length of the first evaporator is halved to 0.405 m, together with halved flat tubes number of 18. Simulation results show a dryness of 0.68 at the outlet, as is shown in Figure 8. Then the two-phase flow of the first evaporator flows into the gas-liquid separator before it enters the compressor. And the remaining 32% liquid refrigerant enters the next evaporator which also has the length of 0.405m. Different number of flat tubes of the second evaporator have been simulated which are 18, 13, 9 and 5 respectively in order to study their impact on heat transfer efficiency under the same inlet air parameters of former experiments, while the CO2 inlet dryness is assumed to be 0 in the second-stage evaporator.

Figure 8. Heat transfer along channel of first-stage evaporator.


Figure 9. CO2 dryness of second-stage evaporator under different flat tubes number.


Figure 10. CO2 temperature of second-stage evaporator under different flat tubes number.

Figure 9 shows that when flat tubes number is decreased to 9, CO2 at the outlet of the second stage evaporator is in the overheat zone with overheat temperature of 6°C as is shown in Figure 10. When the number decreased to 5, it turned into two-phase zone. Considering only the postpone of dry-out point, in order to ensure the over-heated state at the outlet, the number of flat tubes should be 9 for optimization of the secondary evaporator, which can reduce the heat transfer area to 37.5% of the original one. Former heat transfer capacity of the evaporator was 3.46 kW, and after the optimization, it has become 2.48 kW for the first stage evaporator, and 0.65 kW for the second stage evaporator, which adds to 90.5% of the former one.

C. Simulation analysis

In order to further reduce the size of evaporator, we plan to figure out the proper operation conditions in which the outlet CO2 will be in the overheated zone. for evaporator with less flat tubes. Different inlet air temperature, humidity and velocity have been analyzed through simulation for studying their impact on such smaller evaporators as is shown in Table 4. Heat transfer rate in each interval was calculated for securing the location of dry-out point. Figure 11 to 13 showed the impact of inlet air parameters on heat transfer performance:

Table 4. Research conditions.


Air inlet temperature (°C)

Relative humidity (%)

Air face velocity (m/s)

Figure 12




Figure 13




Figure 14





Figure 11. Influence of air temperature on the heat transfer rate.


Figure 12. Influence of air relative humidity on the heat transfer rate.


Figure 13. Influence of air speed on the heat transfer rate.

We can see the same pattern in heat transfer rate which is similar to former device that the dry-out point marks a threshold of a drastic decline of heat transfer rate after it. While under all inlet conditions, total heat transfer rate remains to be about 600 W. Figure 12 and 13 show that when air temperature and humidity increase, dry out happens earlier. But the increase of velocity may also improve total heat transfer rate Moreover, air temperature has more significant effect than humidity on the location of dry-out point. Therefore, in order to create a proper condition for smaller evaporator like this in order to have better heat transfer performance, we found that under air temperature of 28°C, humidity of 40% and velocity of 5 m/s, CO2 can reach the dry-out point in the evaporator with 5 flat tubes with overheated CO2 at the outlet.


Considering dry and wet conditions on air side and different heat transfer characteristics of CO2 in two-phase and overheated region, a distributed parameter simulation model of the CO2 microchannel evaporator was established. The heat transfer in the two-phase region was calculated by Cheng correlation, while in overheated region, it was selected according to the Reynolds number. Comparison between experimental and simulated values in terms of CO2 temperature, wall temperature, inlet and outlet air temperature and convective heat transfer coefficient showed little discrepancy which verifies the simulation. Both inlet parameters impact on heat transfer performance and structural optimization have been realized through simulation process according to which we can dawn the follow conclusions:

1)    The comparison between experimental and simulation results shows little discrepancy within 18% which verifies the simulation method.

2)    Convective heat transfer coefficient reached the maximum at the dry-out point and then decline drastically which causes heat transfer deterioration. In the overheated region, the heat transfer coefficient is way smaller compared to that of the two-phase region. Therefore, the later the dry-out happens, the better the cooling efficiency of the device.

3)    Structural improvement of the evaporator has been made by separating one evaporator into two with gas-liquid separator between them. Results show that 37.5% area of the original experimental device can still achieve 90.5% heat transfer rate of the former one.

4)    Dry out occurs in a 5-tube evaporator when the air temperature is 28°C, with humidity of 40% or air velocity of 5 m/s. Higher air temperature or relative humidity makes the dry out happen earlier, while none of which has apparent impact on the total heat transfer. Higher air velocity not only makes the dry-out occurs earlier, but also improves total heat transfer rate.


This paper has been admitted by CCHVAC in 2018. L.J. F.Y.J and L.G. thank to the committee members of CCHAVC and sponsors of TICA company, China.


[1]    Zhao Y and Ohadi M. M, “Experimental study of supercritical CO2 gas cooling in a microchannel gas cooler”, ASHRAE Trans., vol. 110 (1), pp. 291-300. 2004.

[2]    Zan Wu. “Mechanism and predictive methods for flow boiling in micro/mini-channels and micro-fin tubes.” Ph. D. Thesis. Zhe Jiang University, China, 2013.

[3]    Yoon S H, Cho E S and Hwang Y. W. “Characteristics of evaporative heat transfer and pressure drop of carbon dioxide and correlation development”, International. Journal. Refrigerator.,vol. 27 (2), 2004, pp 111-119.

[4]    Cheng L, Ribatski G and Wojtan L “New flow boiling heat transfer model and flow pattern map for carbon dioxide evaporating inside horizontal tubesInt. J. Heat Mass Transfer, vol. 49 (21-22), 2006, pp 4082-4094.

[5]    L Wojtan, T Ursenbacker, J R “Thome, Investigation of flow boiling in horizontal tubes: part II – development of a new heat transfer model for stratified-wavy, dry-out and mist flow regimes”, Int. J. Heat Mass Transfer, vol. 48(4), 2005, pp2 970-2985.

[6]    Haiqing Zhang, Bei Guo, “Numerical simulation of the carbon dioxide microchannel evaporator using distributed parameter model,” Journal of Xi’an Jiao Tong University, vol. 46 (1) 2012, pp 42-47.

[7]    V. Gnielinski, “New equations for heat and mass transfer in turbulent pipe and channel flow”, Int. Chen. Eng. vol.16 (2) 1976, pp 401-409.

[8]    Sieder E N, Tate G E, Heat transfer and pressure drop of liquids in tubes. Ind Eng Chem., 28 (1936):1429-1435.

[9]    Pettersen J, “Flow vaporization of CO2 in microchannel tubes,” Exp. Therm. Fluid Sci.,vol. 28 (2-3), 2004, pp 111-121.

[10]  Kim M H, Bullard C W, “Air-side thermal hydraulic performance of multi-louvered fin aluminum heat exchangers”, International Journal Refrigerator.,vol.25 (3), 2002, pp 390-400.

Fu Yijun, Li Guo and Lv JingPages 31 - 39

Stay Informed

Follow us on social media accounts to stay up to date with REHVA actualities


0 product in cart.products in cart.