Page 321 - Pipeline Risk Management Manual Ideas, Techniques, and Resources

14/298Absolute Risk Estimates
Table 14.9 Relationship between wall thickness and failure rate frequency and then with various consequence scenarios. A rela-
tive risk model serves an operator especially well when it pro-
Wall thickness (mm) Failure rate (jer IO00 km-year) vides guidance and decision support for resource allocations. It
shows him the system vulnerabilities and points to mitigation
0-5 0.750 measures to remedy them. The consequence assessment is
5-1 0 0.220 mostly there to indicate the priorities and perhaps suggest the
10-15 0.025
appropriate level of mitigation. There are normally few oppor-
Source: European Gas Pipeline Incident Data Group (EGIG) 1993 tunities to significantly change the consequences directly (see
report. Chapter 7).
Consequences are more critical in risk communications and
frequency with which certain hole sizes have been observed for regulatory decision making, often leading to the need for
various failure modes (seeTable 14.7). absolute risk values. This makes a study and quantification of
Reference [67] suggests that some adjustments should be incident event sequences more necessary. Many ofthe events in
applied to recommended failure rates (shown in Table 14.5), in the sequences studied will be related to a particular damage
order to account for observed reductions in failure rates with state. The sequence begins with a failure probability but then
increasing wall thickness. These are shown in Table 14.8 and were follows paths that are ultimately measuring the likelihood of
used in Table 14.6. A relationship between wall thickness and fail- various consequence scenarios-is there immediate ignition or
ure rate due to external forces is also given in Table 14.9. delayed ignition? How big a cloud may form? What are the
Similarly, a relationship to depth of cover is offered inTable 14.10. likely temperature and wind conditions? What if an explosion
Potential risk reduction benefits from several mitigation occurs? How far are the vulnerable receptors?
measures, as suggested by various references, have been com- The overall likelihood of failure of the pipeline-ofien the
piledinTable 14.11. starting point for the event sequence-is a hction of all of the
variables discussed in Chapters 3 through 6 of this book. Most
risk assessment efforts similarly focus on the probability of
IV. Relative to absolute risk failure. This is not only because failure frequency reduction is
usually the best way to reduce risks, but also because so many
As previously noted, it may be advantageous to marry a relative variables impact failure frequency that a model is needed to
assessment of the probability of failure with an absolute failure properly consider all of the important factors.
Inferring a failure frequency from a relative risk score is
Table 14.10 Summary of failure frequencies (per 1000 km-years) by illustrated later in Case Study C. The concept could also be
depth of burial
applied to the other case studies. The process involves a corre-
lation between a failure frequency curve and relative risk
Depth ofpipeline
burial Normal0.9m 1.5m 2m 3m scores. Ideally, this would be established by many data points,
demonstrating that certain failure frequencies are to be
Mechanical failure 0.143 0.143 0.143 0.143 expected with risk scores produced by a certain model. As
Operational 0.047 0.047 0.047 0.047 more and more companies practice formal risk management
and gather data over several years for many miles of pipe, this
Corrosion 0.085 0.085 0.085 0.085 relationship will solidify. Case Study C is forced to make the
Natural 0.013 0.013 0.013 0.013 linkage with only one data point and the end points of the risk
External impact 0.132 0.099 0.066 0.0013 score scale-the minimum three points required to define a
Total 0.42 0.387 0.354 0.289
curve.
Source: Morgan, B., et ai., “An Approach to the Risk Assessment of Case Study C takes advantage of the fact that the end points
Gasoline Pipelines,” presented at Pipeline Reliability Conference, of a relative risk assessment scale also have meaning. A good
Houston, TX. November 1996. scoring model should show that one end of the scale represents
Table 14.1 I Some reported mitigation benefits
Mitigation Impact on risk Reference
Increase soil cover 56% reduction in mechanical damage when soil cover increased from 1 .O to 1.5 m 70
Deeper burial 25% reduction in impact failure frequency for burial at 1.5 m; 50% reduction for 2m; 99% for 3m 58
Increased wall thickness 90% reduction in impact frequency for > I 1.9-mm wall or >9.1 -mm wall with 0.3 safety factor 58
Concrete slab Same effect as pipe wall thickness increase 58
Concrete slab Reduces risk of mechanical damage to “negligible” 70
Underground tape marker 60% reduction in mechanical damage 70
Additional signage 40% reduction in mechanical damage 70
Increased one-call awareness
and response 50% reduction in mechanical damage
Increased ROW patrol 30% reduction in mechanical damage 70
Increased ROW patrol 30% heavy equipment-related damages; 20% rancWfann activities; 10% homeowner activities 86
Improved ROW, signage,
public education 5-15% reduction in third-party damages 86