A fundamental benefit of liquid recirculation is the improvement of the heat transfer coefficient on the refrigerant side of the evaporator. This improvement is attributable to improved wetting of the refrigerant-side surfaces and increasing the velocity of refrigerant. A measure of the extent of overfeed is called the recirculation number or circulation ratio, n

For there to be overfeed of liquid, n must exceed 1.0. The immediate question that arises is what is the optimal value of n? The value of n affects the performance of the evaporator in several ways. First, an increase in n increases the heat-transfer coefficient without limit. But the penalites of increased n are twofold: increased cost of pumping the liquid, and increased pressure drop through the evaporator. The increased refrigerant pressure drop results in a higher evaporating temperature for a given pressure leaving the coil.

The influence of n can be visualized as in Fig. 8.5. The basis of comparison of various values of n should be the combined power of the compressor and liquid pump to provide a given refrigeration duty. Refrigeration duty means the rate of heat transfer at the specified temperature of fluid or product being cooled. The air and refrigerant temperature distributions are profiled throughout the coil in Fig. 8.5. The distribution of the air and refrigerant temperatures are shown for two different values of n.

One value of n is the optimal one that was found by some means, and the other n is higher than optimum. The mean-temperature difference between the air and refrigerant is represented by the area between the curves of the air and refrigerant temperatures. The mean-temperature difference for the high n is less than for the optimum n, because the heat-transfer coefficient is higher with the high n. But the greater pressure gradient with the high n translates to a steeper slope of the evaporating temperature curve, and in this case, the outlet temperature and pressure from the evaporator will be lower with the high n. The result is that a lower compressor suction pressure must be provided which requires more compressor power for a given refrigerating capacity. Not shown in Fig. 8.5 is that the power for the liquid pump will be slightly higher with high n, and the pressure drop in the liquid/vapor return line from the evaporator to the low- pressure receiver will also be higher. These two conditions are further penalties of the high value of n.

The above-mentioned general trends must be verified by laboratory measurements and field experience. Wile1 measured the influence of the circulation ratio in laboratory tests of a steel, finned tube, ammonia evaporator cooling air, and the results are shown in Fig. 8.6. The outside diameter of the tubes of the coil was 16 mm (0.625 in). The overall heat-transfer coefficient is expressed as a ratio to that experienced with the coil operating under control of an expansion valve that maintains a small amount of superheat. Figure 8.6 indicates that a 25% improvement in heat-transfer capacity is possible with a circulation ratio of 3 or higher in comparison to a ratio of 1. Wile’s tests showed little or no improvement when n increased beyond 4 or 5.

The coil in Wile’s tests experienced a pressure drop of 10 kPa (1.5 psi) when n=7 and when the evaporating temperature was –29°C (–20°F). Associated with this pressure drop is a 1.7°C (3°F) higher boiling temperature at the entrance of the evaporator than at the outlet. Wile observed that, while the typical practice is to specify the circulation ratio in selecting the coil, it is the refrigerant flow rate through the coil that more precisely expresses its optimum performance. The same coil could be selected for different capacities, depending upon the airto-refrigerant temperature differences, resulting in widely different refrigerant flow rates for a given n.

Lorentzen conducted some experiments similar to those of Wile, but with the additional parameter of heat flux. As the heat flux (the rate of heat transfer per unit area) increases, the U-value of the coil also increases, as illustrated in Fig. 8.7. This effect of the heat flux is generally observed in boiling heat transfer, at least until the flux is so high that vapor blankets the surface. Lorentzen’s studies showed a progressive increase in the U-value as n increased, in contrast to optimum U-value reached in Wile’s experiments. Figure 8.7 shows an abrupt increase in U-value with the first liquid that overfeeds the coil. Indeed, continued increase in n improves the U-value, but the most important benefit occurs by converting the coil from direct expansion to a value of n that is even slightly greater than 1.0.

Richards3 analyzed the work of several authors including the experimental data of Van Maale and Cosijn4 that is shown in Fig. 8.8, and concluded that to achieve favorable heat-transfer coefficients the Froude number is a distinguishing

characteristic. The Froude number is a dimensionless term reflecting the ratio of the inertia to the gravity forces and in its simplest form is:

A Froude number for liquid, Frlig, is a modification of Fr in Eq. 8.2 and is

The value of Frliq that appears to result in favorable heat-transfer coefficients is 0.04, which is associated with the achievement of annular flow in the tubes. This value of Frliq is noted on Fig. 8.8.

Some other recommendations of the circulation ratio are as follows. One manufacturer of coils5 recommends: n=4 for ammonia, and n=3 for R-22. Geltz recommends a higher n when the coil is circuited for top feed of refrigerant to achieve good wetting of the evaporating surfaces. That proposal is also apparent in the ASHRAE Handbook7 whose recommendations are shown in Table 8.1. Another recommendation reflected in Table 8.1 is that n for the halocarbons can be less than for ammonia. The reason for choosing a lower value of n for halocarbons is also to avoid excessive pumping power. In comparison to ammonia, the latent heat of the halocarbon refrigerants is lower, but the liquid density is higher. These two reasons are partially compensating, but in general, for a given refrigeration capacity the power required by the liquid pump in an ammonia system is about one-third that in a halocarbon system.

While circulation ratios from 2 to 4 are standard, the possibility of choosing a circulation ratio much higher (between 20 and 40) also has been explored on an R-12 plate-type freezer. With conventional values of n, evaporation is likely to occur through the entire evaporator—from entrance to exit. The temperature of refrigerant thus drops in the direction of flow as the pressure drops. With very high values of n, the temperature profile is likely to be as shown in Fig. 8.9, where the liquid is under pressure at the evaporator entrance and absorbs the evaporator load through its increase in sensible heat, reflected in a rise in temperature. Somewhere in the evaporator, perhaps near the outlet, the rising saturation pressure meets the falling actual pressure and vaporization begins. The use of extremely high values of n may have merit, but only in special cases.

that good evaporator performance will result with values of n between 3 and 4, and indeed these have been the traditional design values. In the past few years, however, attention has been directed toward low-charge systems. Systems with low inventories of ammonia are attractive for safety reasons. For halocarbon systems, new CFC replacements are expensive, so the cost of the refrigerant charge becomes a factor. One way to reduce the refrigerant charge in a liquid recirculating system is to reduce the value of n, which results in a greater fraction of vapor in the evaporator coils and the liquid/vapor return line. The density of vapor is less than liquid, so the mass will be less. Furthermore, the low-pressure receiver can be smaller and contain less charge.