Page 32 - Pipeline Risk Management Manual Ideas, Techniques, and Resources

Conceptualizing a risk assessment approach 1/11
mon question asked by the evaluator will be “How do you the mission statement or objective of the risk management
know?” This should be asked in response to almost every program. The underlying reason may vary depending on
assertion by the interviewee(s). Answers will determine the the user, but it is hoped that the common link will be the
uncertainty around the item, and item scoring should reflect desire to create a better understanding of the pipeline and
this uncertainty. This issue is discussed in many of the its risks in order to make improvements in the risk picture.
suggested scoring protocols in subsequent chapters. Secondary reasons or reasons embedded in the general purpose
What defaults are to be used when no information is may include
available? See the discussion on uncertainty in this chapter
and Chapter 2. Identify relative risk hot spots
0 Ensure regulatory compliance
Set insurance rates
What sources of variation exist? Define acceptable risk levels
Prioritize maintenance spending
Typical sources of variation in a pipeline risk assessment 0 Build a resource allocation model
include 0 Assign dollar values to pipeline systems
Track pipelining activities
Differences in the pipeline section environments
Differences in the pipeline section operation Having built a database for risk assessment purposes, some
Differences in the amount of information available on the companies find much use for the information other than risk
pipeline section management. Since the information requirements for compre-
0 Evaluator-to-evaluator variation in information gathering hensive risk assessment are so encompassing, these databases
and interpretation often become a central depository and the best reference source
Day-to-day variation in the way a single evaluator assigns for all pipeline inquiries.
scores
Every measurement has a level of uncertainty associated VI. Conceptualizing a risk assessment
with it. To be precise, a measurement should express this uncer- approach
tainty: 10 fti 1 in., 15.7”F~0.2’.Thisuncertaintyvaluerepre-
sents some of the sources of variations previously listed: Checklist for design
operator effects, instrument effects, day-to-day effects, etc.
These effects are sometimes called measurement “noise” as As the first and arguably the most important step in risk
noted previously in the signal-to-noise discussion. The varia- management, an assessment ofrisk must be performed.
tions that we are trying to measure. the relative pipeline risks. Many decisions will be required in determining arisk assess-
are hopefully much greater than the noise. If the noise level is ment approach. While all decisions do not have to be made dur-
too high relative to the variation of interest, or if the measure- ing initial model design. it is useful to have a rather complete
ment is too insensitive to the variation of interest, the data list of issues available early in the process. This might help to
become less meaningful. Reference [92] provides detailed avoid backtracking in later stages, which can result in signifi-
statistical methods for determining the “usefulness” of the cant nonproductive time and cost. For example, is the risk
measurements. assessment model to be used only as a high-level screening tool
If more than one evaluator is to be used, it is wise to quantify or might it ultimately be used as a stepping stone to a risk
the variation that may exist between the evaluators. This is eas- expressed in absolute terms? The earlier this determination is
ily done by comparing scoring by different evaluators of the made, the more direct will be the path between the model’s
same pipeline section. The repeatability of the evaluator can design and its intended use.
be judged by having her perform multiple scorings of the same The following is a partial list of considerations in the design
section (this should be done without the evaluator’s knowl- of a risk assessment system. Most of these are discussed in
edge that she is repeating a previously performed evaluation). If subsequent paragraphs of this chapter.
these sources of variation are high, steps should be taken to
reduce the variation. These steps may include 1. Purpose-A short, overall mission statement including the
objectives and intent of the risk assessment project.
0 Improved documentation and procedures 2. Audience-Who will see and use the results of the risk
Evaluator training assessment?
Refinement of the assessment technique to remove more General public or special interest groups
subjectivity Local, state, or federal regulators
0 Changes in the information-gathering activity Company-all employees
Use of only one evaluator Company-management only
Company-specific departments only
3. Uses-How will the results be used?
Why are the data being collected? 0 Risk identrficafion-the acquisition of knowledge, such
as levels of integrity threats, failure consequences and
Clearly defining the purpose for collecting the data is impor- overall system risk, to allow for comparison of pipeline
tant. but often overlooked. The purpose should tie back to risk levels and evaluation of risk drivers