Page 396 - Pipeline Risk Management Manual Ideas, Techniques, and Resources

Sample algorithms /371
1. Pipeline rating-prioritizing pipelines for a particular damages, the role ofpipe strength, the potential for internal cor-
maintenance action rosion, and potential consequences (LIF). Words in brackets
2. Failure-mode rating-applying the scheme to one pipeline represent risk variables. Brackets identify variables for some
and determining and prioritizing the most likely causes of versions of SQL software. Variables multiplied by factors are
failure and implementing a maintenance program to prevent normally done to weight the importance or to convert some
these failure causes. measurement (such as deDth of cover in feet) to a Doint score.
,
-
For example, the equation ([depth-cover] = [cover] / 3 * IO)
Due to the lengthy calculations involved, a Microsoft means that a variable called "cover''-a measure of depth of
Windows PC program was developed to compute the relative cover-is divided by 3 and multiplied by a weighting factor of
risk values. Fifty-nine questions are asked of the user that IO to arrive at a point score for the calculated variable called
relate to the design, condition, history, and environment for "depth-cover." In this case, [cover] represents the actual meas-
each pipeline segment. Input and output screens are shown in urement of cover in inches and [depth-cover] is the risk vari-
the paper. able created from that measurement.
Variables are left in their abbreviated form, but should be
readily recognizable by experienced pipeline personnel.
Sample algorithms
The following are examples of algorithms used to assess risk
variables such as the probability of damage from third-party

Variable Interpretation
Sample Third-party Damage Algorithm
ThdF'tySum (([depth-cover]+ [activity]+[exposed-facilities]+ Sum of all subvariables adjusted by leak history
[one-call]+[patrol]+ [public-edn]+[ROW-condl)'
(IIf( [leak]=l,0.9,IIf([leak]=5,0.9,1)))' [repair-thd-pty])
depth-cover ([earth cover] x [earth type]) + [pavement]
activity (l/([utilities]+ [one-calk]+ [pop]'2)*12 + First, as the number of foreign utilities, one-call
[prev activity IevelIR) reports, and/or population density increases,
the activity score decreases (worsens).
Population density is the most important
indicator so it is doubled in this first part of the
calculation. This is then multiplied hy 12, a
scaling factor, and by a previous assessment of
activity level, divided by 2 to reduce its impact
(since it is older information).
exposed-facilities ([ab\._gmd]/2+Ilf( [cover]=0,0,5)) x [vulnerability] A vulnerability assessment (done elsewhere) is
multiplied by the type of exposure. If an
aboveground component has been previously
identified, it will appear as a nonzero value in
the [abv-grnd] variable. As another check. if
depth of cover is zero, then that is also
considered an exposure. If no exposures are
found, maximum points (5 pts) are awarded.
one-call ([mandated] + [effectiveness] + [use] + [response]) Combines variables that evaluate properties of a
one-call system and the company's reaction to
one-call reports.
public-edn ((2 x [door-to-door]) + [mail out] +[advertisement]) x Combines variables that evaluate aspects of a
[pub ed fieq] public education system.
ROW-cond ([undergrowth] + [overgrowth] + [signs/markers]) Combines variables that evaluate aspects of
ROW condition.
patrol ([air_patrl_freq]'Ilf( [ROW]<3 ,O. 5, [air_patrl_effl)*2) Combines variables that evaluate properties of
patrol-frequency and effectiveness.
Miscellaneous Algorithm Variables
pipe-fctr. (Ilf( [pipe-maxpress]/[MOP]< 1 ,0,IIf(( [pipe-maxpress]/
[MOP])>2,20,( [pipe-maxpress]/[MOP]-l)*20))' As a measure of pipe strength, the ratio of
[ILI-design-flaw]) available strength versus operating pressure is
scored (max 20 pts) and adjusted by the results
ofthe most recent search for flaws by in-line
inspection.
Continued