Page 316 - Pipeline Risk Management Manual Ideas, Techniques, and Resources

141293

Absolute Risk

Estimates

duction 14/293

General failure data 141295
Additional failure data 14/2
Relative to absolute risk 14
V. Index sums versus failure pro
diction 141301
obabilitzes 1413
e limits 141304
EX. Receutorvulnerabilities 141305 Ph.rt.
Population 141305
Generalized damage states

1. Introduction also often implies a precision that is usually not available to
any risk assessment method. So, the “absolute scale” offers the
As noted in Chapter 1, risks can be expressed in absolute terms, benefit of comparability with other types of risks, whereas the
for example, “number of fatalities per mile year for permanent “relative scale” offers the advantage of ease of use and cus-
residents within one-half mile of pipeline. . .” Also common is tomization to the specific risk being studied.
the use of relative risk measures, whereby hazards are priori- Note that the two scales are not mutually exclusive. A rela-
tized such that the examiner can distinguish which aspects of tive risk ranking is converted into an absolute scale by equating
the facilities pose more risk than others. The former is a fre- previous accident histories with their respective relative risk
quency-based measure that estimates the probability of a spe- values. This conversion is discussed in section IV on page 298.
cific type of failure consequence. The latter is a comparative Absolute risk estimates are converted into relative numbers by
measure of current risks, in terms of both failure likelihood and simple mathematical relationships.
consequence. Each scale has advantages, and a risk analysis that marries
A criticism of the relative scale is its inability to compare the two approaches may be the best approach. A relative
risks from dissimilar systems-pipelines versus highway trans- assessment of the probability of failure can efficiently capture
portation, for example-and its inability to provide direct fail- the many details that impact this probability. That estimate can
ure predictions. The absolute scale often fails in relying heavily then be used in post-failure event sequences that determine
on historical data, particularly for rare events that are extremely absolute risk values. (Also see Chapter 1 for discussion of
difficult to quantify, and on the unwieldy numbers that often issues such as objectivity and qualitative versus quantitative
generate a negative reaction from the public. The absolute scale risk models.)