Page 317 - Pipeline Risk Management Manual Ideas, Techniques, and Resources

14294 Absolute Risk Estimates
Although risk management can be efficiently practiced Property damages
exclusively on the basis of relative risks, occasionally it Thermal radiation levels
becomes desirable to deal in absolute risks. This chapter pro- Overpressure levels from explosions.
vides some guidance and examples for risk assessments requir- Total consequences expressed in dollars
ing absolute results-risk estimates expressed in fatalities,
injuries, property damages, or some other measure of damage, Ifthe damage state of interest is more than a “stress” level such
in a certain time period-rather than relative results. This as a thermal radiation level or blast overpressure level, then a
requires concepts commonly seen in probabilistic risk assess- hazard area or hazard zone will also need to be defined. The
ments (PRAs), also called numerical risk assessments (NRAs) hazard area is an estimate of the physical distances from the
or quantitative risk assessments (QRAs). These techniques pipeline release that are potentially exposed to the threat. They
have their strengths and weaknesses as dmussed on pages are often based on the “stress” levels just noted and will vary in
23-25, and they are heavily dependent on historical failure size depending on the scenario (product type, hole size, pres-
frequencies. Several sources of failure data are cited and their sure, etc.) and the assumptions (wind, temperature, topography,
data presented in this chapter. In most instances, details of the soil infiltration, etc.). Hazard areas are discussed later in this
assumptions employed and the calculation procedures used to chapter and also in Chapter 7.
generate these data are not provided. Therefore, it is imperative Receptors within the defined hazard area must be character-
that data tables not be used for specific applications unless the ized. All exposure pathways to potential receptors, as discussed
user has determined that such data appropriately reflect that in Chapter 7 should be considered. Population densities, both
application. The user must decide what information may be permanent and transient (vehicle traffic, time-of-day, day-of-
appropriate to use in any particular risk assessment. week, and seasonal considerations, etc.); environmental sensi-
Case studies are also presented to further illustrate possible tivities; property types; land use; and groundwater are some of
approaches to the generation of absolute risk values. This chap- the receptors typically characterized. The receptor’s vulnerabil-
ter therefore becomes a compilation of ideas and data that ity will often be a function of exposure time, which is a function
might be helpful in producing risk estimates in absolute terms. of the receptor’s mobility-that is, its ability to escape the area.
The careful reader may conclude several things about the The event sequences are generated for all permutations of
generation of absolute risk values for pipelines: many parameters. For a hazardous substance pipeline, impor-
tant parameters will generally involve
0 Results are very sensitive to data interpretation.
Results are very sensitive to assumptions. Chance of failure
Much variation is seen in the level of detail of analyses. Chance of failure hole size
0 A consistency of approach is important for a given level of Spill size (considering leak detection and reaction scenarios)
detail of analysis. Chance of immediate ignition
Spill dispersion
Chance of delayed ignition
II. Absolute risks Hazard area size (for each scenario)
Chance of receptor@) being in hazard area
As noted in Chapter 1, any good risk evaluation will require the Chance of various damage states to various receptor.
generation of scenarios to represent all possible event sequences
that lead to all possible damage states (consequences). To esti- A frequency of occurrence must be assigned to the selected
mate the probability of any particular damage state, each event damage state-how often might this potential consequence
in the sequence is assigned a probability. The probabilities can occur? This frequency involves first an estimate of the proba-
be assigned either in absolute terms or, in the case of a relative bility of failure of the pipeline. This is most often derived in part
risk assessment, in relative term-showing which events hap- from historical data as discussed below. Then, given that failure
pen relatively more often than others. In either case, the proba- has occurred, the probability of subsequent, consequence-
bility assigned should be based on all available information. In a influencing events is assessed. This often provides a logical
relative assessment, these event trees are examined and critical breakpoint where the risk analysis can be enhanced by combin-
variables with their relative weighting (based on probabilities) ing a detail-oriented assessment of the relative probability of
are extracted as part of the model design. In a risk assessment failure with an absolute-type consequence assessment that is
expressing results in absolute numbers, the probabilities are sensitive to the potential chains of events.
assigned as part of the evaluation process.
Absolute risk estimates require the predetermination of a
damage state or consequence level of interest. Most common is 111. Failure rates
the use of human fatalities as the consequence measure. Most
risk criteria are also based on fatalities (see page 305) and are Pipeline failure rates are required starting points for determin-
often shown on FN curves (see Figure 14.1 and Figure 15.1) ing absolute risk values. Past failures on the pipeline of interest
where the relationship between event frequency and severity are naturally pertinent. Beyond that, representative data from
(measured by number of fatalities) is shown. Other options for other pipelines are sought. Failure rates are commonly derived
consequence measures include from historical failure rates of similar pipelines in similar
environments. That derivation is by no means a straightfor-
Humaninjuries ward exercise. In most cases, the evaluator must first find a
Environmental damages general pipeline failure database and then make assumptions