A.4 Risk Tolerability

Once a scenario is modelled and a percentage likelihood of fatality is calculated, it is necessary to determine if that risk is acceptable. This section discusses the issue of acceptable risk.

A.4.1 The concept of risk and its definition

The terms risk and hazard should be differentiated.

Hazard is a physical situation with a potential for human injury, damage to property, damage to the environment, or some combination of these.

Risk is the likelihood of a specified undesired event occurring within a specified period. It may either be a frequency (the number of specified events in a given period) or a probability (the chance of the specified event following a prior event). Mathematically, risk is a function combining both the failure events and the consequences of them.

Risk occurs in every human activity and is virtually impossible to eliminate totally without avoiding the activity completely. As a generality for industry, if the activity is to take place at all, the risk should be kept ALARP, and the remaining risk has in any event to be at a level that is acceptable to workers in the workplace, the public at large and, the appropriate regulatory authorities or a standard internal to an organisation. These authorities will usually define the criteria for acceptability of any risk as well as policing compliance with measures necessary to manage the risk at the levels intended.

A.4.2 Introduction to the elements of risk

Some hazards by their nature result only in a risk to individuals (i.e. one person is affected at any one time). It can be appropriate to express the level of risk simply in terms of likelihood of death in a year. Examples of such measures would be:

- The risk of death by falling from a ladder or scaffolding108.

- The risk of being fatally struck by lightning in the UK (per year)109 is 1 in 10 000 000.

- The risk of being killed by a meteor impact large enough to produce a 1 km2 crater110 is 1 in 768 000 000.

Comparison of levels of risk by this method is generally unsatisfactory, because the likelihood of exposure of the affected population is seldom clear. Such measures can, however, be useful in establishing the tolerability criteria111 to be applied to risk since the example is a generally unavoidable (involuntary) risk taken without question by the general public.

Better measures of individual risk include statistics based on death (or serious injury) per unit of activity. This takes into consideration the exposure time of the individual to the hazardous activity. The UK chemical industry developed the statistic of fatal accident rate (FAR) that has been extended to cover a wide range of industrial and other activities.

The FAR is the number of deaths expected per 100 million exposed hours. Put more descriptively, the number of deaths expected in a workforce of 1 000 during a working lifetime.

A commonly published example of this measure is:

- The FAR for the Chemical and Allied Industries in the period 1987-1990 was 1,2112

- The FAR for a commercial airliner in 2006 was 0,6113

Again there can be difficulty in using and comparing statistics on this basis. In this example, the scope of the data source, The Chemical and Allied Industries is not disclosed. Given that the great proportion of fatal accidents within the chemical industry are not particularly related to the process hazards is not shown; trips, falls from heights and contact with objects predominate. The chemical industry can show progressive improvements in the FAR over time, due to safety management techniques. However, the publication of such data in the literature lags far behind this reduction, and can affect views on risk criteria acceptability.

QRA calculations for a particular hazardous process or activity can generally produce risk results expressed or illustrated in the following terms:

- Individual risk.

- FN curves.

Both of these could be relevant to the hazard analysis of a CO2 separation, compression and distribution system, and are detailed in the following sections. Risk contours and a prediction of societal risk are also an output from assessments, but these are not relevant to offshore installations.

A.4.3 Individual risk

The IChemE (Jones) definition for individual risk is the frequency at which an individual may be expected to sustain a given level of harm, from specified hazards. Such a risk is location-specific (e.g. on a specific platform), and also is dependent on the fraction of time a person is likely to be at each hazard location in question. Assuming the 'level of harm' is defined as fatality, the individual risk is equivalent to the FAR term defined above, except that it will usually be expressed as fatalities per year, rather than per hour of exposure.

A.4.4 FN curves

The total risk associated with a pipeline or facility is obtained by calculating the value for each consequence, and then adding all the individual risk values together. The result of this can be expressed as an FN curve as shown in Figure A.2. F is the cumulative frequency of fatalities (or other serious events), N being the consequence term (usually expressed as numbers (N) fatalities). Because the values of F and N typically extend across several orders of magnitude, both axes on an FN curve are frequently logarithmic. Plots of the results of various alternative risk reduction strategies can easily be placed on the same graph. After plotting a known tolerability limit on the diagram, decisions on the acceptability of the predicted risk may be made.

A hypothetical FN curve114 along with illustrative tolerability limits (yellow and green lines) is shown as Figure A.2.

Figure A.2 Example FN curve

108 The odds of serious risks that people can relate to, 2002, http://riskcomm.com/visualaids/riskscale/datasources.pho

109 Process Safety Analysis – An Introduction, Skelton, R., 1997 (IChemE, Rugby UK)

110 Damage by impact, the case at Meteor Crater, Arizona, LMV Martel, Hawaii Institute of Geophysics and Planetology, 11 December, 1997

111 The basis for establishing risk criteria is not fully discussed here. General references on the subject such as Lees' Loss Prevention in the Process Industries should be consulted.

112 Mannon S. (ed), Lees' Loss Prevention in the Process Industries 3rd edition 2005 (Elsevier Butterworth – Heinemann, Oxford UK)

113 Helicopter Safety in the Oil and Gas Business, B. Tender, Shell Aircraft International, May 2006

114 Managing Risk DNV