Natural refrigerants

NATURAL REFRIGERANTS
FOR INDUSTRIAL REFRIGERATION

A B Pearson CEng
Star Refrigeration Ltd

There are five refrigerants recognised as “natural”; ammonia, carbon dioxide, water, air and the group of short-chain hydrocarbons (ethane, ethylene, propane, propylene, butane and isobutane).  Ammonia is in common use as a refrigerant, accounting about 15% of the refrigerant market, and nearly all of this use is in the industrial sector.  The other natural refrigerants are hardly used at all in this sector, apart from some specialist systems using hydrocarbons, for example in petrochemical plants.  In recent years (since about 1998) however, several systems have been installed which use carbon dioxide as a refrigerant.  This marks a significant change in this market.

Before the first round of fluorocarbon phase out (1986-1995) the industrial market made substantial use of CFCs and HCFCs, particularly R22 and R502 and particularly in the United Kingdom and France.  Ammonia retained a greater market share in other parts of the world, particularly Eastern Europe, Russia and the United States.  It seems reasonable to assume that the use of ammonia in these markets would have declined in favour of fluorocarbons as a result of stricter health and safety requirements, if it had not been for the Montreal Protocol.

In the early stages of the CFC phaseout some end-users who would previously have used R502 switched to R22 and others accepted ammonia as the most suitable refrigerant.  Within a few years however it became apparent that R22 was not likely to remain in use in the longterm and the market shifted towards ammonia.  For most countries this was a minor issue, as they had widespread experience of ammonia and a sympathetic regulatory background.  In the United Kingdom there was a bigger shift, but it was accomplished relatively easily.  In France the same shift has not occurred so easily, owing to stricter and more complex regulations governing the use of ammonia.

HFCs have not been adopted as industrial refrigerants to any great extent anywhere in the world.  There is no single component replacement for R502 – the most promising, R125, has a very low critical point and so tends to be inefficient.  The most promising blend, R404A, is expensive, requires an expensive lubricant which does not tolerate moisture, and tends to leak from large, site installed systems.  Some other blends have further disadvantages, such as temperature glide, which make them unsuitable for use in flooded systems.  The cost of the refrigerant should not be under-estimated in this failure of HFCs to dominate the industrial market.  For the first time in the 150 year history of industrial refrigeration the cost of the refrigerant became a mathematically significant proportion of the capital cost of the installation.

Many end-users have accepted the apparent disadvantages of ammonia, its toxicity and flammability, with reluctance.  It is recognised as the most cost-effective option for industrial plant in cold stores, freezers and process plant – even with the costs of regulatory compliance.  In a few cases however, this reluctance has prompted system designers to consider other natural refrigerants.  Water is inappropriate for low temperature plant and air systems are far too inefficient to be considered for most applications.  Hydrocarbons are generally perceived as too dangerous for use in large charge industrial systems and so have not received serious consideration.  Several designers around the world however have chosen to adopt carbon dioxide as refrigerant for industrial plant.  It is in fact the only non-toxic, non-flammable, non-ozone depleting, non-global warming refrigerant suitable for use in traditional Rankine cycle refrigeration plant.

Carbon dioxide has now been used in this way for industrial systems in the United Kingdom, France, Germany, the Netherlands, Switzerland, Australia, Japan and the United States of America.  These systems fall into two principal types – those which use a carbon dioxide compressor and those which evaporate and condense the carbon dioxide at nominally the same pressure, using the fluid as a “volatile secondary” refrigerant.  The majority of systems are of the latter type, usually with an ammonia plant cooling the carbon dioxide condenser.  These are cost effective where brine or glycol was the alternative because of the greatly reduced pumping and pipework costs.  Until recently the former type has been much less common due to the lack of suitable compressors, but recent systems have used modified versions of standard recip or screw compressors adapted to cope with discharge pressures in the range 30-40 Bar.

Current screw compressors are not able to provide gas at sufficiently high pressure for hot-gas defrosting, and the reciprocating compressors which have been adapted for this use are not yet on general release to the market.  The following table gives an overview of a few of the industrial systems recently installed with details of the type of system and type of defrost.

Table 1 – Some recent carbon dioxide installations

Notes: FD = freeze drier, PF = plate freezer, T = tunnel, S = spiral

The last of these installations is significant because it is a composite chill and cold storage distribution centre serving a supermarket chain, and therefore it combines a compressed carbon dioxide system for the cold store with a pumped carbon dioxide system for the chill chambers.  In a system of this type and size hot gas defrosting was essential but there was no suitable compressor on the market.  The defrost method used takes liquid at the chill condition (-5C saturated pressure) and pressurises it to about 9C saturated pressure, about 45 Bar(G).  This liquid is pumped through a heat exchanger with warm glycol from the ammonia plant’s oil cooling circuit.  The resultant vapour is slightly superheated in a second heat exchanger, connected in series on the glycol side, and the resultant gas, at 45 Bar(G) and 20C is supplied to the coolers for defrosting.  This defrost system is potentially more efficient than a traditional ammonia hot gas system, as it is not necessary to run the ammonia compressors at an artificially high discharge pressure to achieve defrost.  The only extra power consumed is by the carbon dioxide high pressure pump, which is very small.  A schematic diagram of the refrigeration system is shown in Figure 1 and the defrost system is shown in Figure 2.

Figure 1 – Schematic diagram of Carbon Dioxide and Ammonia systems

The alternative for this end-user would have been a pumped glycol system for the chill store and a separate low charge ammonia system for the cold store.  The installed system was nominally the same cost to install, but is cheaper to run; the carbon dioxide pumps have 4kW motors, whereas the equivalent glycol pumps would have required 55kW motors.  The annual energy saving is estimated to be about £20,000 (30,000 Euros).  It is also expected that maintenance costs for the carbon dioxide system will be slightly lower than for the equivalent glycol plant.

Figure 2 Schematic diagram of HP liquid defrost system

There is no doubt that the capital cost and running efficiency of this type of system will be improved over time as new products are introduced to the market and system design concepts are refined.  At present the majority of industrial refrigeration systems use ammonia as refrigerant and for end-users willing to accept the hazards associated with direct ammonia this will remain the most cost-effective approach in the near future.  For those who cannot accept direct ammonia a carbon dioxide cascade system, either pumped or compressed depending on operational requirements, now provides a suitable alternative to direct ammonia, and is already more cost effective for large installations than a fluorocarbon system or an indirect glycol system.

Future developments will include the proliferation of components for use in the range 40-55 Bar(G) in the near future, increasing the options for defrosting.  In the longer term the next generation of compressors will offer the option of avoiding ammonia completely, if they can provide discharge pressures of 90 Bar(G) or more.  This would enable large system design based upon the transcritical cycle to be considered, and in turn would introduce the possibility of carbon dioxide in large high temperature systems such as office air-conditioning and district cooling.  At present the simple transcritical cycle is not efficient enough, except where high grade heat recovery, for example in water heating, is required.  There is therefore also a need for system design development to find more efficient ways of implementing carbon dioxide systems.  It is interesting to note that the same techniques would provide ways of making higher pressure fluorocarbons such as R125 or R410A more attractive for large systems, although the problems of cost, lubricants, tolerance of water and system tightness might still tip the balance in favour of carbon dioxide.