Page 201 - Pipeline Risk Management Manual Ideas, Techniques, and Resources

8/178 Data Management and Analyses
set of algorithms used to define risk. To many, the term model Step 4: assessing risks
refers to the software that holds data and presents risk results.
This is actually the environment in which the model resides, not After a risk model has been selected and the data prepared, risks
the model itself, by the terminology of this book. along the pipeline route can be assessed. The previously
The environment housing and supporting the risk model is selected risk assessment model can be applied to each segment
critical to the success of the risk management process. to get a unique risk score for that segment. These relative risk
numbers can later be converted into cumulative risk values
and/or absolute risk numbers.
111. Risk management process Moving from overviews of risk down to the smallest details
of a specific piece ofpipe will often be necessary. Software that
This section describes the mechanics of performing a risk supports rapid tabularization and perhaps map overlays is use-
assessment using the common software tools of a spreadsheet ful. Spill flowpath or dispersion modeling, soil penetration, and
and desktop database. The risk assessment process is designed hazard area determinations are aspects of the more robust risk
to capture pertinent information in a format that can be used to assessments.
first create segments with constant risk characteristics and then
assign risk scores to those segments. Step 5: managing risks
Many of the data processing steps in a risk assessment
can appear complex when first studied, especially when those Having performed a risk assessment for the segmented
steps are described rather than demonstrated. Most processes pipeline, now comes the critical step of managing the risks.
are, however, fairly self-evident once the risk assessment In this area, the emphasis is on decision support-providingthe
efforts are under way. As such, the reader is advised not to tools needed by the risk manager to best optimize resource
become reluctant to embark on the effort based on the appar- allocation.
ently significant issues of the process, but rather to begin the This process generally involves steps such as these:
effort and use the following sections as a reference document
as issues arise. Analyzing data (graphically and with tables and simple sta-
Chapter 1 presents an overall process for risk management. tistics)
This process can be revisited when considering potential soft- 0 Calculating cumulative risks and trends
ware environments since the software will ideally fully support Creating an overall risk management strategy
each step in the process. 0 Identifying mitigation projects
Performing “what-if” scenarios.
Step 1: risk modeling
Patterns, trends, and relationships among data sets can
As previously noted, a pipeline risk assessment model is a set become an important part of managing risks. Software that
of algorithms or “rules” that use available information and supports analytical graphics routines will be useful.
data relationships to measure levels of risk along a pipeline.
A model can be selected from some existing and commercially Definitions
available models, customized from existing models, or created
“from scratch” depending on your requirements. Algorithms Several terms in this discussion might be used in a manner that
can be created to use data directly from a database environment is unfamiliar to the reader. Terminology is not consistent among
to calculate risk scores. Several common software environ- all risk modelers so these definitions are more for convenience
ments will support efficient data storage, retrieval, and algo- in describing subsequent steps here.
rithm calculations. Each record must have an identifier that relates that record to
a specific portion of the overall pipeline system, that is, an ID.
Step 2: data preparation This identifier, along with a beginning station and ending
station, uniquely identifies a specific point on the pipeline sys-
Data preparation or conditioning produces data sets that tem. It is important that the identifier-stationing combination
are ready to be loaded into and used by the risk assessment does indeed locate one and only one point on the system. An
model. Data preparation includes processes to smooth or alphanumeric identification system, perhaps related to the
enhance data into zones of influence, categories, or bands as pipeline’s name, geographic position, line size, or other com-
may be appropriate. Computer routines greatly facilitate these mon identifying characteristics, is sometimes used to increase
processes. the utility of the ID field.
Risk variables are also commonly called events in keeping
Step 3: segmentation with modem GIS terminology. The current characteristics of
each event are called conditions (also sometimes called atm’b-
Because risks are rarely constant along apipeline, it is advanta- Utes or codes). For example, for the event (population), possible
geous to first segment the line into sections with constant risk conditions include residential high, residential low, commer-
characteristics (dynamic segmentation) or otherwise divide cial high, etc. For the event mapshecords, possible conditions
the pipeline into manageable pieces. This might be a one-time are excellent, fair. poor. none. The event depth of cover could
or rare event, to ensure consistent segments. Alternatively, it have a number or numerical ranges such as 24”, 19”, > 48”, or
might change every time the underlying data change. Again, 12-24” as its possible conditions. Events, as variables in the
computer routines facilitate this. risk assessment, can be named using standardized labels.