EFCTC: Accidents with refrigeration fluids

Due to number of accidents with ammonia or hydrocarbons, it is necessary to look in detail at the risks associated with their use. This aspect is even more important for developing countries, because of situations like uneven maintenance, unreliable power supplies, climate conditions /Ref 1998 - ch. 12.7/.

Although sound safety training and controls might help, the extra cost needed to ensure a high health and safety level for using toxic or highly explosive material (in some cases plants have been totally destroyed) should also be taken into account.

A Working Party on Refrigerant System Safety has been set up by the International Institute For Refrigeration (IIR). Indeed, although some refrigerants have been in use for 100 years, very little documented experience exists, which makes safety assessment an issue, and therefore safety standard a challenge. IIR also considers the interest for correlating existing safety requests with experienced accidents. The Working Party aim is to document the expertise gathered by experienced engineers and to record it in an objective way. For all types of systems and all refrigerants the real number of accidents should be investigated, and a database be set up for accidents involving injuries to humans and property damages.


Accidents in refrigeration applications

The use of hydrocarbons and ammonia is often considered as safe, provided stringent safety measures and standards are followed. Such statements are based on risks assessments (example /Ref 1999/ Ch. 9.2.4 about hydrocarbons in road transport). However, these risk assessments are only theoretical. An IEA study /Berg 1994/ concluded that a VROM (NL) ammonia calculated risk assessment lead to a risk of deaths "10-100 times smaller than those observed.

Accidents involving ammonia

There are several reports on accidents involving ammonia /CDCIR 1992/ /Berg 1994/ Amm 1995/ which conclude that it is difficult to collect data on all the accidents that occur. Unep 1998 compiled a list of significant accidents which contains 18 ammonia accidents.

Amm 2002/ makes a comprehensive analysis of more than 400 accidents involving ammonia between 1992 and 2001, wherefrom 109 involved ammonia. Most (but not all) of the accidents involving “Other refrigerants” actually were the consequence of a fire, rather than a leakage.

"Working Fluid Safety" /Berg 1994/ has used statistics on ammonia refrigeration accidents to derive probabilities of ammonia accidents. We have applied these probabilities to a Dutch proposal /Eco 1998/ to switch to ammonia 70 000 refrigeration systems by 2010 (instead of 150 today). The results presented in Table 5 are striking - for a population of only 15 million.

Kind of riskFrequencyNL Probability (70 000 units)
Death 1 x 10-4 7 per year
Injury 10-30 x 10-4 70 - 210 per year
Accident 10-150 x 10-4 70 – 1 050 per year
Ammonia Risk Assessment applied to the Netherlands /Berg 94/ /Eco 98/

These risks are specially valid for smaller units, because large ammonia refrigeration units are usually surveyed 24 hours a day. For the large number of users of smaller refrigeration units, this will most often not be the case, since many units are left unattended for long periods of time.

Due to millions of refrigeration equipment world-wide including developing countries, we think that lives and environment will be better protected by not introducing ammonia outside its traditional uses.

Accidents involving Hydrocarbons

Since refrigeration accidents are due to equipment failure, poor maintenance or human factors, there is no reason to expect different accident frequencies, except that the toxicity risk would be replaced by fire and explosion risk of a hydrocarbon leakage, which can have a significant destructive force.

Again the number of installations should be kept in mind : the same Dutch report /Eco 98/ suggests switching to hydrocarbons by 2010 a number of 295.000 units containing 1 to 30 kg refrigerant, of 10.000 units containing more than 50 kg. Bearing in mind that the explosion of 1 kg hydrocarbon releases the same energy as 1 kg of TNT, we wonder if this change should not be promoted with more care for human or property losses.

The report /IIR 97/ concludes that "Expanding the use of flammable refrigerants [..] changes past safety practices and should be done with great care to ensure that the safety of the public is maintained."

Accidents involving Fluorocarbons

There is also little published data for risks related to HFCs, but some is available on use of CFCs and HCFCs. These have led to some fatalities over the years, mainly due to suffocation. This is a risk with any inert gas; e.g. engineers have died in confined areas that are full of refrigerant vapour. The number of fatalities reported is far lower than for ammonia systems, even though there are significantly more FC systems in use. The risk of fatalities with FCs is several orders of magnitude lower than for ammonia (Berg 1994).