Conclusions and Recommendations

While the Log Mean Temperature Difference is often used to compute the size (UA) of a heat exchanger, this calculation becomes invalid for mixtures when the working fluid exhibits a nonlinear temperature glide. An analysis of the derivation of the LMTD reveals assumptions that must be discarded. Most notably, it can no longer be assumed that the temperature will be an easily-determined function of the heat exchanged.

In the case of ammonia-water, which was selected for study due to its strong nonlinearity, calculations utilizing the LMTD resulted in dramatic errors. For small pinch points, these errors may lead to heat exchangers that are undersized by as much as a factor of fifty. The LMTD is simply not valid, and a numerically calculated mean temperature should be used instead. For a given pinch point, the error generally increases with the ammonia concentration. At large pinch points, however, the magnitude of the differences in the error decreases between concentrations.

For the refrigerant mixtures proposed as replacements for HCFC-22, the error is less, but clearly still exists. The magnitude of the error does not appear to directly correlate with the size of the temperature glide, as can be observed from comparing refrigerants 1 and 2 at a pinch of 2 K. Both their operating pressures and their glides are nearly the same, but refrigerant 1 has a standard error of 1.34% while refrigerant 2 has a SE of 0.01%.

In general, the UA error decreases as the size of the pinch increases for both NH₃/H₂O and the replacement refrigerants. However, this is not the case for refrigerant 2; its error increases until a pinch of 4 K, and at that point begins to decrease. For all other refrigerants, though, the change in the error also decreases with increasing pinch, which is both mathematically and conceptually logical.

Modified cycles to exploit the unique characteristics of mixtures should be analyzed using the correct mean temperature difference. The potential efficiency improvement could then be properly evaluated for these refrigerant mixtures. Since the majority of refrigerants being considered in applications currently satisfied by R-22 are zeotropes, both the desorber-absorber heat exchange (DAHX) cycle and the liquid-line/suction-line heat exchanger (LLSL-HX) should be considered for new systems and for retrofits.

A DAHX cycle incorporates both thermal and mechanical controls to optimize the utility of a temperature glide. Thermal control is achieved through heat exchange between the evaporator and the condenser, while mechanical control is gained through the variable speed liquid pump. This desorber-absorber heat exchange can provide the same temperature lift as a conventional cycle, but over a smaller pressure lift (Didion & Bivens, 1990). So by definition, the working fluid of a DAHX cycle must be a mixture whose constituents have very different boiling points; this means that a large temperature range can be covered at constant pressure through composition shifting (Radermacher, 1987). The DAHX cycle also uses a liquid-line heat exchanger with the evaporating zeotrope. An expansion device is used instead of a turbine in the compression cycle, but adding a liquid-line heat exchanger recovers the lost work.

The LLSL-HX cycle subcools the high-pressure refrigerant by reducing the superheating of the vapor entering the compressor. An LLSL-HX can either improve or worsen a system's performance, depending on what working fluid is used. For R-407c, for example, employing a LLSL-HX with a counterflow evaporator and condenser for cooling raises the efficiency to 5-6% above that of R-22, even when the same alterations are made to the R-22 cycle (Bivens et al., 1995). The impact of LLSL-HX on the volumetric capacity can be even greater (Domanski, Didion & Doyle, 1994).

In addition to considering new ways to utilize synthetic refrigerants, the advantages of switching to natural refrigerants should also be considered. While CFCs and HCFCs are currently designated as harmful to the environment, they were considered to be safe, and even ideal, for many years. It is unknown which synthetic refrigerants that are now thought of as benign will prove dangerous in the future. Some studies suggest, in fact, that HFC-134a may be decomposed in the troposphere by sunlight and form acid and poisonous substances (Lorentzen, 1995). On the other hand, natural refrigerants such as ammonia-water, propane, and carbon dioxide have long existed in the biosphere. Furthermore, while refrigerants such as R-410a and R-407c have zero ozone depletion potential (ODP), they do have non-trivial global warming potential (GWP). NH₃/H₂O, propane, and CO₂ also have zero ODP, and while it is obvious that CO₂ has a high GWP, the GWP of both propane and NH₃/H₂O is negligible.

Throughout the past, the greatest concern with the use of ammonia has been safety. While it is true that ammonia can burn with air and is poisonous, current refrigeration standards and ammonia's other natural properties may offset these difficulties. Ammonia gas is much lighter than air, and can easily be vented away. If it is trapped, its strong smell serves as a warning that there is a leak. In fact, ammonia can be detected at concentrations as low as five parts per million in air, before there is any significant danger. In terms of flammability, the lower ignition limit of ammonia is three to seven times that of natural gas and most hydrocarbons. The pressure of ammonia was also once considered a limiting factor, but compared to the much higher pressures of the synthetic refrigerants in this study, that should no longer be counted as a concern. Even disregarding these natural safety features, the need to create systems that are tight and stay tight lessens the danger of ammonia as a refrigerant (Lorentzen, 1995).

Ammonia as a refrigerant has been applied to both industrial and residential heating and cooling units. It has been used in water chillers and heat pumps for residential buildings; direct-expansion systems for supermarket refrigeration; large heat pump plants for district heating; and central cooling of gas turbine cogeneration plants (Cavallini, 1996). Ammonia is also becoming more acceptable to the public. While more popular in Europe, especially Scandinavia and Germany, ammonia has the long-term potential to increase its U.S. market share as ``green'' issues gain in importance. Other natural refrigerants are also becoming more prominent: When a natural refrigerant-powered home refrigerator (marketed under the name ``GreenFreeze'') was recently introduced in Germany, the manufacturer sold out within weeks. So, in summary, while systems can be developed that will increase the feasibility of synthetic refrigerants as R-22 replacements, the opportunity to change existing systems should perhaps be redirected toward an effort to reduce the use of all synthetics in favor of natural replacements.