Manuscript ID: 326

DOI: 10.18462/iir.icr.2019.326

Stefano FILIPPINI, Umberto MERLO, Giovanni MARIANI, Livio PERROTTA,

LU-VE GROUP, Uboldo, 21040, Italy, [email protected]

ABSTRACT

CO2 is often proposed as a radical solution to eliminate the greenhouse effect caused by halogenated hydrocarbons belonging to the HFC category. Even though the direct contribution of CO2 is practically zero, the indirect effect would be increased if the CO2 refrigeration cycles were less efficient than traditional ones (lower COP), due to larger electricity consumption bringing about larger emissions of CO2 and of other pollutants from power stations. The appropriate choice of heat exchanger technology is a fundamental condition for obtaining COP values form CO2 cycles allowing for a real reduction of the greenhouse effect. However, CO2 is significantly different from all the other halogenated and non-halogenated fluids and it poses peculiar opportunities to heat exchanger designers: their discussion is the subject of this paper, underlining in particular innovative solutions and test laboratory to enhance the heat exchange efficiency and to improve the overall performance of the refrigeration plant.

Keywords: CO2, unit coolers, gas coolers, heat exchangers, COP, high efficiency

1.    INTRODUCTION

In several refrigeration applications, CO2 is an excellent and environmentally-friendly solution; LU-VE has designed very efficient heat exchangers for CO2 applications and has been able to make important developments thanks to the use of its modern test ring. This plant can test the performance of CO2 finned heat exchangers, both air cooled unit coolers and gas coolers. The new testing plant enabled the launch of a specific project for a CO2 fin-and-tube heat exchanger, with the primary aim of improving knowledge of heat exchange phenomena in evaporation, condensation and during trans-critical gas cooling. The influence of oil on internal heat exchange coefficient also enters into the scope of the research. The paper describes the testing activity, the calibration made on software that calculates product performance, and potential improvements to products. In particular, it was possible to calibrate a specific method able to take into account the behaviour of the fluid during trans-critical cooling, properly considering all the parameters affecting real performance.

2.    TEST PLANT

The CO2 plant (Fig. 1) was designed to be able to test condensers, gas coolers and unit coolers. The maximum operating pressure is 120 bar, while the maximum temperature is 120°C. The capacities of the plant in the standard conditions specified below are:

• Unit Coolers (SC2): 15 kW
• Gas Coolers e Condensers (100bar, 120°C, DTmin=3°C) e (25/40/65°C): 40

A dedicated software programme was developed in-house using LabView to monitor and acquire data. The tests were conducted making a thermal balance between the test unit and a contrast group in order to guarantee the reliability of the experimental data. This required a great deal of time for the calibration of the measurement instrumentation and in particular for the definition of dispersion inside the climatic test chamber.

Figure 1: Image of data acquisition codes and view of the plant

3.    CO2 HEAT EXCHANGERS

In refrigeration plants using CO2 as the working fluid, two types of heat exchangers are used:

• Evaporators, which are included in every proposed plant configuration: in direct CO2 cycles, in binary cycles (using a low temperature CO2 cycle and an higher temperature cycle, operated by another fluid and rejecting heat towards the ambient) and in other systems using CO2 as the cold energy carrier, condensed by a refrigerating machine and evaporated by the users device. Evaporators, working at low temperature, do not require elevated operating pressure and therefore are not substantially different from models for halogenated fluids.
• Gas-coolers, which are included in direct cycles only to reject heat towards the ambient. They perform the same duty of conventional fluid condensers, but rather than condensation (implying a two-phase equilibrium) a simple transition from the expanded gas phase to the liquid state takes place. As a matter of fact, as CO2 has a very low critical temperature of 31°C, a supercritical operating pressure is necessary to maintain a temperature higher than the one of ambient receiving heat from the cycle. The critical pressure being 73.8 bar, operating pressures much larger than those of conventional cycles will be adopted.
• The typical shape of supercritical cycles is shown in 2. Compared to conventional cycle rejecting most of their thermal capacity at a constant temperature, supercritical cycle performance is not only influenced by the minimum and maximum pressure, but their COP is strongly affected by the gas cooler outlet temperature, i.e. the temperature of the liquid at the expansion device entrance¹. This is very important to obtain acceptable COP: as a matter of fact, CO2 cycles perform very well with low coolant temperatures (e.g. water-heating heat pumps, low ambient temperatures in cool regions). For a given ambient temperature, the gas cooler exit temperature is imposed by the design characteristics of the gas cooler, therefore assuming a fundamental role as far as the cycle performance are concerned.

Figure 2: Examples of inverse CO2 supercritical cycles.

4.    EVAPORATORS

A CO2 evaporator for refrigeration applications does not have to undergo especially high working pressures (Table 1). However it is necessary to prevent overpressures caused by prolonged standstill of the equipment or by defrosting, when the temperature can rise well over that of the cooling room. Rather than oversizing the evaporator and the refrigerant lines, it is preferable to adopt expedients which can limit the project pressure to 60 bar or even less (safety valves, pump-down to remove liquid from the evaporator). Such pressure values are just a little above those normally used in refrigeration (all LU-VE evaporators are available up to 85 bar) and do not impose any special design, even if larger thickness of coil tubes and headers are usually adopted.

Table 1: Relationship between temperature [°C] and pressure [bar] for CO2.

On the other hand, it is interesting to determine if an evaporator designed for conventional refrigerants can operate correctly for CO2, with no or limited modifications, and, if so, to estimate the variations of thermal power. It should be stated in advance that the thermophysical properties of CO2 are favourable to obtaining elevated heat transfer performance. Compared to R404A, CO2 has higher specific heat, higher thermal conductivity and lower viscosity. This last fact, along with the greater vapour density, allows fewer pressure drops at the same mass velocity. Considering that (at equal capacity) the larger heat of evaporation brings about a lower throughflow, pressure drop reductions at the same power turn out to be very significant indeed. Table 2 shows the results of a theoretical prediction of a LU-VE unit cooler running on CO2 (in terms relative to R404A) at two different evaporation temperatures, in the following hypotheses;

• Unchanged specifications: a slight increase in power at -8°C, becoming more consistent at low temperatures (from 3.5 to 11%); fluid velocity and pressure drops are very low.
• Reducing the number of feeds: in-tube velocity return to optimal values and 6-7% capacity improvement is shown compared to the previous case; reducing the number of feedings reduces the cost of the gas header and distributor.
• Reducing the number of inlets and using smooth tubes instead of microfin tubes (helically grooved microfins such as those normally used in LU-VE unit coolers): microfin tubes are particularly useful with poor refrigerant heat transfer coefficient: their convenience is very reduced at high evaporation temperature, but remains significant at low temperature with a low density fluid (-30°C).

Table 2: Comparative performance of unit coolers for R404A and CO2. The ratios are valid for some representative models but are not applicable in general.

The last two solutions permit a modest improvement of the specific cost (€/kW) of the equipment, as long as the pressures of the design do not exceed 40-60 bar. LU-VE has already supplied various clients with CO2 unit coolers up to now no visible indications have arisen of the slightest power deficit nor of any operating problems.

5.    CO2 HEAT EXCHANGERS

The gas cooler design is notably more complex, also due to the larger operating pressure (up to 150 bar), and poses some relevant peculiarities. The fundamental aspect for the thermodynamic design is that, as a consequence of the high average temperature along the upper isobar (responsible for the modest COP values), with CO2 it is possible to bring the cooling air to much higher temperatures than those occurring with a refrigerant having a condensation phase at constant temperature. Fig. 3 shows this situation very clearly: it is evident that with CO2 an air ∆T 2-3 times greater can be obtained. Consequently it is possible to use an airflow reduced by the same proportion at equal thermal capacity. The large reduction in the airflow gives notable advantages in terms of reduced front area of the fin pack, of electric power required for ventilation and of the initial cost of the fans and their regulators.

To quantify these statements, a calculation method was developed capable of accounting for the particular distribution of the ∆Ts between CO2 and air (as in Fig. 3), provided that flows are arranged to run countercurrent².

Figure 3: Heat transfer diagram for a CO2 gas cooler and for a condenser using a conventional refrigerant.

The exchanger is subdivided into 20 computational sections: for each one an independent evaluation is done of the average logarithmic ∆T and of the in-tube heat transfer coefficient, with the Gnielinski correlation for single phase flows. Fig. 4 shows an example of how some important parameters vary in the computational sections. It can be noticed that: (i) the heat transfer coefficient presents a maximum close to the critical point , (ii) the required surface area increases significantly in the cold end, due to the reduced ∆T between the two fluids and to the low liquid velocity.

Figure 4:Variations of some parameters in the computational sections of a CO2. gas cooler

Table 3 shows a comparison between a R404A condenser (capacity of about 170kW with initial ∆T of 15K) and CO2 gas coolers of the same power range. Since the CO2 outlet temperature plays a preponderant role, the comparison was carried out in two ways: (i) at equal power, varying the final temperature, and (ii) at a final ∆T of 5 K, varying the power. The number of parallel feedings is optimised in all cases. The following solutions are proposed in table 3:

• The first solution is the R404A reference (in normal production).
• The second solution presents the same fin dimensions (frontal area and rows) and the same ventilation. The rating is exuberant (last line) or, as an alternative, a very reduced

∆T can be obtained (the 0.3 value is, however, only valid for perfect counterflow) all of which is caused by the very large ∆T between CO2 and air (at equal air flow). The above mentioned possibility of reducing the airflow was not exploited in this solution

• The third solution thoroughly exploits this possibility, using only one fan instead of
• The exchanger surface is redistributed to best adapt to a reduced airflow: the number of rows is doubled and the front section was halved, with a heat transfer surface practically the same as the original. The thermal rating at final ∆T of 3K is slightly less than the reference (-4%) in the presence of major reductions in the dimensions (50%), in the ventilation power (66%) and in the noise level (4.8dB). It should be pointed out that in these cases the outlet air temperature is in the range of 60°C: it is therefore convenient to place the fans at the coil inlet (forced draft) to avoid thermal stress to the motor and to increases the mass air flow, compared to the usual solution of induced draft (fan at coil outlet).

Table 3: Comparative performances of air cooled condensers with R404A and CO2 under the following conditions: air temperature 25°C, condensation R404A 40°C, CO2 pressure 100 bar.

In general, the optimum solutions may vary depending on the design survey and on the requirements imposed by the compatibility with existing models, for industrial reasons. However, one can conclude that the use of CO2 could bring about significant reductions in the size of the equipment (in relation to the reduced ventilation) compared to equipment with similar ratings for conventional refrigerants, even with small final ∆T values (for example, 3K as in table 3).

LU-VE has over 13 years of experience in CO2 ventilated heat exchangers and it participated in the construction of the first trans-critical supermarket in Europe (2004), providing gas coolers with adiabatic spray system in trans-critical operation.

The CO2 gas cooler product can therefore be considered “proven technology” in the refrigeration field. This achievement was made possible because of the design strategy adopted by LUVE, consisting of the utilization of high performance heat transfer surfaces and of miniaturized geometries (small diameter tubes) even for large heat exchangers.

Figure 5: COOP WETTINGEN – Zurig – Switzerland Gas cooler CO2.SHVDT 696 CO2 (2004)

6.    EMERITUS

EMERITUS® is the latest innovation developed for the range of condensers, dry coolers and gas coolers manufactured by LU-VE Exchangers. This new technological advance (patent pending) brings together the benefits of spray systems and adiabatic pre-cooling. A sophisticated control system maximizes the effects of these combined systems.

The specific features of this product make it especially suitable for use in air conditioning and refrigeration. When EMERITUS® is applied to CO2 gas coolers, high system COP can be reached even during the hottest hours of the year. This means that the “CO2 equator” can be redrawn, extending the geographical area where trans-critical systems can be cost-effectively constructed.

Figure 6: Emeritus® technology

The ability to use just the adiabatic system in the absence of a water treatment plant is a further example of its versatility. The graph shows the performance levels of a gas cooler with Adiabatic Panel technology compared with a traditional dry system. The technology with adiabatic system shows clearly superior performance. At equal capacity, it is possible to bring the CO2 outlet temperature down from 40°C to 30°C. A reduction of 10 K of the CO2 has a positive consequent impact on the COP of the system. In a comparison made at equal conditions (the same capacity, the same evaporation temperature, etc.), a COP increase of 69%* was obtained (*hypothesis of a simple cycle with evaporation temperature of -8°C. The COP goes from 1.31 (temp. Tout CO2 gas at 40°C) to 2.21 (T.out CO2 gas at 30°C)).

Figure 7: Comparison of gas cooler capacity ADIABATIC SYSTEM vs DRY

Figure 8: EMERITUS® 2017 – France – Gas cooler with adiabatic system

7.    CONCLUSIONS

The applications of CO2 in the refrigeration industry could shortly become an important reality. From the heat exchanger point of view, the utilization of CO2 poses some problems (greater operating pressures) but also offers notable opportunities, especially in the most difficult design case of the gas coolers. We have seen that reductions of the airflow and of the coil front area can be achieved, at equal capacity and with very small final ∆T values (this last being an essential parameter for obtaining a good COP of the cycle). It brings about lower fan consumption, smaller size and some production cost savings, counterbalanced by the increased use of copper resulting from the thicker tube walls and headers. The fin-and-tube geometries used for conventional fluids are perfectly adequate to CO2 application, in the case of LU-VE production which has for many years concentrated on small diameter tubes even for large units.

¹ Liquid temperature also affects the performance of conventional cycles, provided that a specific heat transfer section is devoted to sub-cooling, but at a much lower extent.

² In plate-fin coils with 3-4 rows (or more) it is usually possible to arrange the circuiting in order to obtain a fluid path very close to counterflow, with negligible influence on the predicted performance.