Page 29 - Pipeline Risk Management Manual Ideas, Techniques, and Resources

1/8 Risk: Theory and Application
compensating for poor visibility by slowing down demon- The experience factor and the intuition of experts should not
strates a simple application of risk management. The driver be discounted merely because they cannot be easily quantified.
knows that a change in the weather variable of visibility impacts Normally little disagreement will exist among knowledgeable
the risk because her reaction times will be reduced. Reducing persons when risk contributors and risk reducers are evaluated.
vehicle speed compensates for the reduced reaction time. If differences arise that cannot be resolved, the risk evaluator
While this example appears obvious, reaching this conclusion can have each opinion quantified and then produce a compiled
without some mental model of risk would be difficult. value to use in the assessment.
Risk management, for the purposes of this book, is the set of When knowledge is incomplete and opinion, experience,
actions adopted to control risk. It entails a process of first intuition, and other unquantifiable resources are used, the
assessing a level of risk associated with a facility and then assessment of risk becomes at least partially subjective. As it
preparing and executing an action plan to address current and turns out, knowledge is always incomplete and some aspect of
future risks. The assimilation of complex data and the subse- judgment will always be needed for a complete assessment.
quent integration of sometimes competing risk reduction and Hence, subjectivity is found in any and all risk assessment
profit goals are at the heart of any debate about how best to methodologies.
manage pipeline risks. Decision making is the core of risk man- Humans tend to have bias and experts are not immune from
agement. Many challenging questions are implied in risk this. Knowledge of possible bias is the first step toward mini-
management: mizing it. One source [88] identifies many types of bias and
heuristic assumptions that are related to learning based on
Where and when should resources be applied? experiment or observation. These are shown inTable 1.2.
How much urgency should be attached to any specific risk
mitigation?
Should only the worst segments be addressed first? 111. Uncertainty
Should resources be diverted from less risky segments in
order to better mitigate risks in higher risk areas? As noted previously, risk assessment is a measuring process.
How much will risk change if we do nothing differently? Like all measuring systems, measurement error and uncer-
tainty arise as a result of the limitations of the measuring tool,
An appropriate risk mitigation strategy might involve risk the process oftaking the measurement, and the person perform-
reductions for very specific areas or, alternatively, improving ing the measurement. Pipeline risk assessment is also the com-
the risk situation in general for long stretches of pipeline. Note pilation of many other measurements (depth of cover, wall
also that a risk reduction project may impact many variables for thickness, pipe-to-soil voltages, pressure, etc.) and hence
a few segments or, alternatively, might impact a few variables absorbs all of those measurement uncertainties. It makes use
but for many segments. of engineering and scientific models (stress formulas, vapor
Although the process of pipeline risk management does not dispersion and thermal effects modeling, etc.) that also have
have to be complex, it can incorporate some very sophisticated accompanying errors and uncertainties. In the use of past fail-
engineering and statistical concepts. ure rate information, additional uncertainty results from small
A good risk assessment process leads the user directly into sample sizes and comparability, as discussed previously.
risk management by highlighting specific actions that can Further adding to the uncertainty is the fact that the thing
reduce risks. Risk mitigation plans are often developed using being measured is constantly changing. It is perhaps useful to
“what-if” scenarios in the risk assessment. view a pipeline system, including its operating environment, as
The intention is not to make risk disappear. If we make any a complex entity with behavior similar to that seen in dynamic
risk disappear, we will likely have sacrificed some other aspect or chaotic systems. Here the term chaotic is being used in its
of our lifestyles that we probably don’t want to give up. As an scientific meaning (chaos theory) rather than implying a disor-
analogy, we can eliminate highway fatalities, but are we really ganized or random nature in the conventional sense of the
ready to give up our cars? Risks can be minimized however-at word. In science, dynamic or chaotic systems refer to the many
least to the extent that no unacceptable risks remain. systems in our world that do not behave in strictly predictable
or linear fashions. They are not completely deterministic nor
Experts completely random, and things never happen in exactly the
same way. A pipeline, with its infinite combinations of histori-
The term experts as it is used here refers to people most knowl- cal, environmental, structural, operational, and maintenance
edgeable in the subject matter. An expert is not restricted to a parameters, can be expected to behave as a so-called dynamic
scientist or other technical person. The greatest expertise for a system-perhaps establishing patterns over time, but never
specific pipeline system probably lies with the workforce repetition. As such, we recognize that, as one possible outcome
that has operated and maintained that system for many years. of the process of pipelining, the risk of pipeline failure is
The experience and intuition of the entire workforce should sensitive to immeasurable or unknowable initial conditions.
be tapped as much as is practical when performing a risk In essence, we are trying to find differences in risk out of
assessment. all the many sources of variation inherent in a system that
Experts bring to the assessment a body of knowledge that places a man-made structure in a complex and ever-changing
goes beyond statistical data. Experts will discount some data environment. Recall the earlier discussion on signal-to-noise
that do not adequately represent the scenario being judged. considerations in risk assessment.
Similarly, they will extrapolate from dissimilar situations that In more practical terms, we can identify all of the threats to
may have better data available. the pipeline. We understand the mechanisms underlying the