Page 133 - Pipeline Risk Management Manual Ideas, Techniques, and Resources

511 10 Design Index
relative volume of metal losses from corrosion, either internal or When an ILI score indicates actual corrosion damage, but
external. The volume of each metal loss indication is approxi- risk assessment scores for corrosion potential do not indicate a
mated by assuming a parabolic shape for the metal loss configu- high potential, then a conflict may exist between the direct and
ration. As an adjustment to risk scores, this variable reduces the the indirect evidence. It will sometimes not be known exactly
previously calculated corrosion index by up to 50%. where the inconsistency lies until complete investigations are
performed. The conflict could reflect an overly optimistic
Crack This variable represents the relative quantity and assessment of effectiveness of mitigation measures (coatings,
severity of cracking and crack-like indications. As an adjust- CP, etc.) or it could reflect an underestimate of the harshness of
ment to risk scores, this variable reduces the previously calcu- the environment. It could also be old damage from corrosion
lated safety factor by up to 90%, recognizing the relative that has since been mitigated.
unpredictability and severity of cracking. To ensure that the corrosion scores always reflect the best
available information, limitations could be placed on corrosion
Pipe wall flaws This is the combination ofthe other four vari- scores, in proportion to the ILI scores, pending results of the
ables described above. As an adjustment to risk scores, this full investigation. This is illustrated in Table 4.1 1 of Chapter 4.
variable reduces the previously calculated safe@ factor by up to
90%, in addition to previous reductions.
E. Land movements (15% weighting in example
After this analysis, each pipeline segment has been charac- model)
terized in terms of the five defect-type variables shown above.
Those five variables each impact a previously determined risk A pipeline may be subjected to stresses due to land movements
score, as noted, In other words, the pipeline segment is penal- and/or geotechnical events of various kinds. These movements
ized for having damages that are evidence of inadequate corro- may be sudden and catastrophic or they may be long-term
sion control, weakened pipe wall, etc. The amount of the deformations that induce stresses on the pipeline over a period
penalty is proportional to the ILI score and the scale maximum of years. These can cause immediate failures or add consider-
possible value of the risk variable. This worst case penalty is set able stresses to the pipeline and should be carefully considered
on the basis of how much influence that factor could have on in a risk analysis.
failure probability. A common categorization of failure causes is external
Default values are set for missing information, usually due to forces. This category blends several failure causes and makes it
a lack of inspection information. Therefore, the default value difficult to separate land movements from third-party damages
represents a condition where no current inspection information as a root cause of the failure. Since this separation is critical in
is available and the presence of some level of anomalies will be risk management efforts, this risk assessment model isolates
conservatively assumed. land movements as a specific failure mode under the Design
In this particular application, it was conservatively assumed Index. The land movement threat is very location specific.
that an ILI yields no useful information after 5 years from the Many miles ofpipeline are located in regions where potentially
inspection date. ILI scores will therefore be assumed to worsen damaging land movements are virtually impossible. On the
by 20% each year until the default value is reached. other hand, land movements are the primary cause of failures,
The ILI score is improved through visual inspections and outweighing all other failure modes, for other pipelines. All of
removal of any damages present. A pipeline segment that is par- these issues make the assignment of a weighting difficult. It
tially replaced or repaired will show an improvement under this often becomes an issue of model scope, as discussed in
scoring protocol since the anomaly count will have been Chapter 2. The suggested weighting presented here should be
reduced which reduces the corresponding defect penalty. The examined in consideration of all pipelines to be assessed with
penalty can also be reduced even if the ILI score does not the risk model. Where land movements are a very high threat, a
improve by anomaly removal. This can happen if a root cause fifth failure probability index can be established specifically
analysis of the ILI anomalies concludes that active mechanisms for the land movement failure mode.
are not present, despite a poor ILI score. For example, the root Land movement or geotechnical issues in general, can be
cause analysis might use previous ILI results to demonstrate categorized in various ways. One method is proposed whereby
that corrosion damage is old and corrosion has been halted. such events are referred to as natural hazards and categorized
This is a rather complex approach and is not fully detailed as shown in Table 5.6.
here. It is included to demonstrate one possible method to more In the following paragraphs, land movements are examined
fully consider evidence from previous ILI in a general (not as the potential for landslides, soil movements, tsunamis, seis-
anomaly-specific) manner. mic events, aseismic faulting, scour and erosion. Additional
threats such as sand dune formation and movement or iceberg
Approach 2 scour (see Chapter 12) can be included in an existing category
or evaluated independently. Land movements specific to the
Another example of an IL1 scoring application where corrosion offshore environment are discussed in Chapter 12.
evaluations are adjusted by recent ILI results is presented here.
First an ILI score is generated that characterizes the overall Landslide
corrosion metal loss in the pipeline segment. This characteriza-
tion could be based on a system similar to that ofApproach 1 or Many of the potentially dangerous land movement scenarios
it could simply involve a scale for accumulating frequency and have a slope involved (Figure 5.5). The presence of a slope adds
severity of wall loss damages in a segment. the force of gravity. Landslides, rockslides, mudflows, and