Page 60 - Pipeline Risk Management Manual Ideas, Techniques, and Resources

Examples of scoring algorithms 2/39
All failure modes are considered. Optimize spending
All risk elements are considered and the most critical Strengthen project evaluation
ones are included. Determine project prioritization
Failure modes are considered independently as well as Determine resource allocation
in aggregate. Ensure regulatory compliance
All available information is being appropriately utilized.
Provisions exist for regular updates of information. VI. Examples of scoring algorithms
including new types of data.
Consequence factors are separable from probability Sample relative risk model
factors.
Weightings, or other methods to recognize relative The relative risk assessment model outlined in Chapters 3
importance of factors, are established. through 7 is designed to be a simple and straightforward
The rationale behind weightings is well documented pipeline risk assessment model that focuses on potential conse-
and consistent. quences to public safety and environment preservation. It pro-
A sensitivity analysis has been performed. vides a framework to ensure that all critical aspects of risk
The model reacts appropriately to failures ofany type. are captured. Figure 2.4 shows a flowchart of this model.
Risk elements are combined appropriately (“and” ver- This framework is flexible enough to accommodate any level
sus “or” combinations). of detail and data availability. For most variables. a sample
Steps are taken to ensure consistency of evaluation. point-scoring scheme is presented. In many cases, alternative
Risk assessment results form a reasonable statistical scoring schemes are also shown. Additional risk assessment
distribution (outliers?). examples can be found in the case studies of Chapter 14 and in
There is adequate discrimination in the measured Appendix E.
results (signal-to-noise ratio). The pipeline risk picture is examined in two general parts.
Comparisons can be made against fixed or floating The first part is a detailed itemization and relative weighting of
standards or benchmarks. all reasonably foreseeable events that may lead to the failure of
V11. Project completion: a pipeline: “What can go wrong?” and “How likely is it to go
Pipeline: Finalize manuals, complete training, ensure wrong?. This highlights operational and design options that
maintenance protocols are in place, and turn system can change the probability of failure (Chapters 3 through 6).
over to operations. The second part is an analysis of potential consequences if a
Risk assessment: Carefully document the risk assess- failure should occur. This addresses the potential consequences
ment process and all subprocesses. especially the should failure occur (Chapter 7). The two general parts corre-
detailed workings of the algorithm or central model. spond to the two factors used in the most commonly accepted
Set up administrative processes to support an ongoing definition of risk:
program Ensure that control documents cover the
details of all aspects of a good administrative program, Risk = (event likelihood) x (event consequence)
including:
Defining roles and responsibilities The failure potential component is further broken into four
Performance monitoring and feedback indexes (see Figure 2.4). The indexes roughly correspond to
Process procedures categories of reported pipeline accident failures. That is, each
Management of change index reflects a general area to which, historically, pipeline
Communication protocols accidents have been attributed. By considering each variable in
each index, the evaluator arrives at a numerical value for that
Study the results index. The four index values are then summed to a total value
(called the index sum) representing the overall failure probabil-
This might seem obvious, but it is surprising how many own- ity (or survival probability) for the segment evaluated. The indi-
ers really do not appreciate what they have available after vidual variable values, not just the total index score, are
completing a thorough risk assessment. Remember that your preserved however, for detailed analysis later.
final risk numbers should be completely meaningful in a The primary focus ofthe probability part ofthe assessment is
practical. real-world sense. They should represent everything the potential for a particular failure mechanism to be active.
you know about that piece of pipe (or other system compo- This is subtly different from the likelihood of failure.
nent)-all of the collective years of experience of your organ- Especially in the case of a time-dependent mechanism such as
ization, all the statistical data you can gather, all your gut corrosion. fatigue, or slow earth movements, the time to failure
feelings, all your sophisticated engineering calculations. If is related to factors beyond the presence of a failure mecha-
you can’t really believe your numbers. something is wrong nism. These include the resistance of the pipe material, the
with the model. When, through careful evaluation and much aggressiveness of the failure mechanism, and the time of
experience, you can really believe the numbers, you will find exposure. These, in turn, can be furtherexamined. For instance.
many ways to use them that you perhaps did not foresee. They the material resistance is a function of material strength;
can be used to dimensions, most notably pipe wall thickness; and the stress
level. The additional aspects leading to a time-to-fail estimate
0 Design an operating discipline are usually more appropriately considered in specific inves-
0 Assist in route selection tigations.