Page 309 - Pipeline Risk Management Manual Ideas, Techniques, and Resources

13/286 Stations and Surface Facilities
IX. Example of risk management mode and determine that some resources should be allocated to
application certain risk reduction actions. Specifically, they want to reduce
the risk of “human error” and “design issues’’ type failures on
Tank farm operator AST Inc. has performed a basic risk assess- three tanks at the Metropolis Station (see Table 13.13). This
ment for all of their facilities. They now have risk scores for will target “overfill” scenarios and other possible failures that
each station and for each tank within each station. They also involve aspects ofhuman error and design issues.
have risk numbers for sumps, pumps, piping, loadingiunload- Operator AST Inc. drills deeper into their risk data to see why
ing facilities, and other equipment groupings. The risk scores risks are greater for these tanks and to see where their risk miti-
represent all available information regarding the facility to gation efforts could be best applied. Because each failure cate-
which it applies. They are readily compared to a statistic or gory is comprised of many risk variables, they can retrieve
some measure of acceptability, as shown in Table 13.12 for a those variables to see why the risk level is too high. They see
sample oftheir data from the Metropolis Station. The risk score that the risk variables listed in Table 13.14 are seen to be weak,
is a summary number that can be broken into failure categories relative to other tanks and company standards.
of external forces, corrosion, human error, and design issues as Operator AST Inc. can view the risk components of likeli-
well as a “consequence-of-failure’’ value (Table 13.13). hood and consequences separately. They see that there are more
(and cheaper) possible actions to prevent-r reduce the likeli-
Risk score = (likelihood) x (consequence) hood of-an event compared with impacting the consequences.
Likelihood = P1+ P2 + P3 + P4 Most of their alternatives in better controlling an event after it
occurs (consequence reducers) are very expensive. Some
and, for example,
immediately rejected consequence-limiting actions include the
following:
P2 = f(product corrosivity, atmospheric conditions, soil resistivity,
moisture content, pipe-to-soil voltages, inspection procedures, liners,
coatings, interference potential, inhibitors, anodes. etc.} 0 Changing product type (less flammable, less persistent in
environment, lower energy content, less toxic, etc.)
AST Inc. has evaluated their data carefully. They determine 0 Changing the receptors (move the station, move the nearby
which tanks pose the greatest risks, which tanks have the town, etc.).
greater likelihood of failure, and which have the greater conse-
quences, should failure occur. They analyze their data by failure Other consequence-reduction possibilities that are more
practical include emergency response, increased leak detection
capabilities, fire suppression systems, better secondary con-
Table 13.12 Summary of relative risk assessment results tainment, and others. Whereas all options can be investigated,
AST Inc. chooses to concentrate for now only on the secondary
Deviation from average containment alternative for consequence reduction.
Equipment tag Riskscore (or ‘hcceptable ’7 (?A) Noting which risk variables are relatively weak also points
directly to what corrective actions can be applied. From pre-
Tank 101 154 -14.0
Tank315 146 -12.5 established project lists and cost data, the operator assesses the
Tank 655 235 -28.1 costs of several mitigative actions. They compare these costs
with the benefit-the risk reduction-predicted by their model,
Table 13.13 Breakdown of summary riskscores
Equipment tag Risk Score Likelihood Consequence Pl-external forces P2-Corrosion P3-Design issues P4-Human error
Tank 101 154 77 2.0 22 19 25 11
Tank315 146 76 1.9 21 24 22 9
Tank 655 235 49 4.8 14 17 16 2

Table 13.14 Evaluation of risk variables
Deviationfrom average (or
Risk variable “acceptable” risk) (96) Notes
Consequence receptors (forTank 655 only) -32 Higher risks due to proximity to population center, water
intakes, and predicted rangeability of spill (flowing river nearby)
Tank level alarms -8 HHA (high-high alarm) only alarms locally-panel light
in office flashes
Staffing levels -2 Once per week visits currently
Personnel training 4 No formal training for loaders-pamphlet only
Secondary containment -1 1 Dikes in need of repair, too permeable, not sufficient volume for
large releases