Page 79 - Pipeline Risk Management Manual Ideas, Techniques, and Resources

3/58 Third-party Damage Index
process works correctly to interrupt a potential failure specific, so frequencies shown are very rough estimates. It
sequence. It is assumed that 60 percent of heavy equipment seems reasonable to assume that activity involving heavy
operators would have knowledge of and experience with the equipment requires more staging, is of a longer duration, and
one-call process and would therefore utilize it. It is further leaves more evidence of the activity. All of these promote the
assumed that the one-call process “works” 80 percent of the opportunity for detection by patrol.
time it is used. (Both assumptions are thought to conservatively Statistical theory confirms that, with a few reasonable
underestimate the actual effectiveness.) This yields a 48 per- assumptions, the probability of detection is directly propor-
cent chance (60 percent x 80 percent) that this variable inter- tional to the frequency of patrols. For example, calculations
rupts the sequence for that type of activity. It is assumed that indicate that the probability of detection in two patrols is twice
one in ten potentially damaging events would be similarly inter- the probability of detection in one patrol if detection of the
rupted in the case of typical homeowner or farmerhancher same event cannot occur in both patrols. This condition is
activity. This is lower than for the heavy equipment operators essentially satisfied for these purposes since patrol sightings
since the latter group is thought to be more targeted with train- subsequent to the initial sighting are no longer considered to be
ing, advertising, and presentations from owners of buried utili- “detections.” The key point here is that the probability that one
ties. The interruption rates reflect improvements over one-call or more events will occur is the sum of their individual proba-
effectiveness at the time period of the incidents, approximately bilities ifthe events are mutually exclusive.
1969 to 1995, which includes periods when there was either no Discounting patrol errors, as the patrol interval approaches 0
one-call system available or it was available but not mandated. hours (a continuous observation of the ROW), the detection
The continuously increasing acceptance of the one-call proto- probability approaches 100 percent. The patrol interval is
cols by the public and the response of the pipeline operator to changing from a historical maximum interval between patrols
notifications combine to create this estimated interruption rate. of 336 hours (once every two weeks on average, although it
Columns 4,5, and 6 examine the possibility that, given that could be as high as three weeks or 504 hours). The mitigation
an activity has escapedthe one-call process, the impending fail- plan requires a patrol every 24,60, or 168 hours, depending on
ure sequence will be interrupted by improved ROW condition, the location. In theory, this improves the detection probability
signs, or public/contractor education. Assumptions of likeli- by multiples of 2 to 14. On the table of activities, patrol inter-
hood range from five in 100 to 15 in 100, respectively. This vals of 24,60, and 168 hours suggest detections of 93 percent,
means that out of every group of threatening activities, at least a 75 percent, and 36 percent of activities, respectively. This
few will be interrupted by someone noticing the ROW and/or a means that, with a maximum interval between patrols of 24
sign or having been briefed on pipeline issues and reacting hours, only 7 percent of activities would go undetected, given
appropriately. In the interest of conservatism, relatively small the assumed distribution of activities. Obviously, the real situa-
interruption rates are assigned to the proposed improvements in tion is much more complex than this simple analysis, but the
these variables although they can realistically prevent an inci- rationale provides a background for making estimates of patrol
dent in numerous credible scenarios. benefits.
Column 7 examines the effect of depth of cover. One refer- In order to make conservative estimates (possibly underesti-
ence [Ref [58] in this book] cites Western European data mating the patrol benefits), the increased detection probabili-
(CONCAWE) which suggests that approximately 15 percent ties under the proposed mitigation plan are assumed to be: 30
fewer third-party damage failures occur with each foot of cover percent, 10 percent, and 20 percent for heavy equipment,
over the normal (0.9 meters). Using this, a length-weighted homeowner, and ranchifarm operations, respectively. This
average depth of cover was calculated for the pipeline, respec- means that about one-third of heavy equipment operations,
tively. The pipeline shows between 7 percent and 4 percent one in every ten homeowner activities, and one in every
improvement, based on the lengths that are covered deeper than five ranchifarm activities would be detected before damage
about 0.9 meters. Based on this, a value of 5 percent was occurred or, in the case of no immediate leak, would provide the
assigned to the cover variable for the “heavy equipment opera- operator time to detect andrepair damages before a leak occurs.
tions’’ type of activity. This means that five out of every 100 Homeowner and ranchifarm actions are judged to be more dif-
potentially damaging third-party activities would be prevented ficult to detect by patrol because such activities tend to appear
from causing damage by an extra amount of cover. For home- with less warning and are often of shorter duration than the
owner activities, depth of cover is judged to be a more effective heavy equipment operations.
deterrent, preventing three out of ten potential damages. One Table 2 converts Table 1 columns 3 through 8 into probabili-
out of ten potentially threatening ranchedfarmer activities are ties of the sequence NOT being interrupted-the “opposite” of
assumed to be rendered non-threatening by depth of cover. Table 1,
Finally, the impact of patrolling is examined in column 8. Column 9 of Table 2 estimates the fraction of times that the
A table of common third-party activities is presented against a line is under enough stress that, in conjunction with powerful
continuum of opportunity to detect, expressed in days (see enough equipment, a rupture would occur immediately. This
patrol figure in Table 1). The “opportunity” includes an esti- stress level is a function of many variables, but it is conserva-
mate of how long after the activity occurs its presence can still tively estimated that 50 percent of the line is under a relatively
be detected. Since third-party activities can cause damages that high stress level. For the 50 percent of the line that could be
do not immediately lead to failure, this ability to inspect evi- damaged, but not to the extent that immediate leakage occurs,
dence of recent activity is important. The table is intended to the mitigation plan’s corrosion control and integrity reverifica-
provide an estimate of the types of activities that can reasonably tion processes, which specifically factor in third-party damage
be detected in a timely manner by a patrol. The frequency of the potential in determining reinspection intervals, are designed to
various types of activities will be very location- and time- detect and remediate such damages before leaks occur.