Page 47 - Pipeline Risk Management Manual Ideas, Techniques, and Resources

2/26 Risk Assessment Process
Table 2.2 Uncertainty and risk assessment
~~
Action Inspection results Risk relevance
Timely and comprehensive inspection performed No risk issues identified Least risk
Timely and comprehensive inspection performed Some risk issues or indications of flaw potential identified: root
cause analysis and proper follow-up to mitigate risk More risk
No timely and comprehensive inspection performed High uncertainty regarding risk issues
Timely and comprehensive inspection performed Some nsk issues or Indications of flaw potential identified-
uncertain reactions, uncertain mitigation of risk Most risk

data; in general, however, worst-case conditions are conserva- uator with consideration given to data costs and desired accu-
tively used for default values. racy. The idea is for each pipeline section to be unique, from a
Uncertainty also arises in using the risk assessment model risk perspective, from its neighbors. So, within a pipeline sec-
since there are inaccuracies inherent in any measuring tool. tion, we recognize no differences in risk, from beginning to
A signal-to-noise ratio analogy is a useful way to look at the end. Each foot ofpipe is the same as any other foot, as far as we
tool and highlights precautions in its use. This is discussed in know from our data. But we know that the neighboring sections
Chapter 1. do differ in at least one risk variable. It might he a change in
pipe specification (wall thickness. diameter, etc.), soil condi-
Sectioning or segmenting the pipeline tions (pH, moisture, etc.), population, or any of dozens of other
risk variables, but at least one aspect is different from section to
It is generally recognized that, unlike most other facilities that section. Section length is not important as long as characteris-
undergo a risk assessment, a pipeline usually does not have a tics remain constant. There is no reason to subdivide a 10-mile
constant hazard potential over its entire length. As conditions section of pipe if no real risk changes occur within those
along the line’s route change, so too does the risk picture. 10 miles.
Because the risk picture is not constant, it is efficient to This type of sectioning is sometimes called dynamic seg-
examine a long pipeline in shorter sections. The risk evaluator nienfution. It can be done very efficiently using modern com-
must decide on a strategy for creating these sections in order to puters. It can also be done manually, of course, and the manual
obtain an accurate risk picture. Each section will have its own process might be suitable for setting up a high-level screening
risk assessment results. Breaking the line into many short sec- assessment.
tions increases the accuracy of the assessment for each section,
hut may result in higher costs of data collection, handling, and Manually establishing sections
maintenance (although higher costs are rarely an issue with
modern computing capabilities). Longer sections (fewer in With today’s common computing environments, there is
number) on the other hand, may reduce data costs but also really no reason to follow the relatively inefficient option of
reduce accuracy, because average or worst case characteristics manually establishing pipeline sections. However. envisioning
must govern if conditions change within the section. the manual process of segmentation might be helpful for
obtaining a better understanding of the concept.
Fixed-length approach The evaluator should first scan Chapters 3 through 7 of this
text to get a feel for the types ofconditions that make up the risk
A fixed-length method of sectioning, based on rules such as picture. He should note those conditions that are most variable
“every mile” or “between pump stations” or “between block in the pipeline system being studied and rank those items with
valves,” is often proposed. While such an approach may be ini- regard to magnitude of change and frequency of change. This
tially appealing (perhaps for reasons of consistency with exist- ranking will be rather subjective and perhaps incomplete, but it
ing accounting or personnel systems), it will usually reduce will serve as a good starting point for sectioning the line(s). An
accuracy and increase costs. Inappropriate and unnecessary example of a short list ofprioritized conditions is as follows:
break points that are chosen limit the model’s usefulness and
hide risk hot spots if conditions are averaged in the section, or 1. Population density
risks will be exaggerated if worst case conditions are used for 2. Soil conditions
the entire length. It will also interfere with an otherwise effi- 3. Coating condition
cient ability of the risk model to identify risk mitigation proj- 4. Age ofpipeline
ects. Many pipeline projects are done in very specific locations,
as is appropriate. The risk of such specific locations is often lost In this example, the evaluator(s) foresees the most significant
under a fixed-length sectioning scheme. changes along the pipeline route to be population density, fol-
lowed by varying soil conditions, then coating condition, and
Dynamic segmentation approach pipeline age. This list was designed for an aging 60-mile
pipeline in Louisiana that passes close to several rural commu-
The most appropriate method for sectioning the pipeline is to nities and alternating between marshlands (clay) and sandy soil
insert a break point wherever significant risk changes occur. conditions. Furthermore, the coating is in various states ofdete-
A significant condition change must be determined by the eval- rioration (maybe roughly corresponding to the changing soil