Page 153 - Reliability and Maintainability of In service Pipelines

Case Studies on the Application of Structural Reliability 139

the nominal value can reduce the lifetime of the pipeline system by more than
19 years (Fig. 5.9). This is expected to an extent, given that the thinner the pipe
wall, the sooner depth of corrosion reaches the wall thickness of the pipe.
Consequently, should there be any defects along the pipe wall an increased depth
of corrosion is likely to lead to failure. Regardless of the limit state it is being
designed for, the impact of wall thickness upon the system failure of the pipeline
is significant (Fig. 5.9). Generally, choosing a wall thickness that is changed from
its nominal value to either a higher or lower value drastically changes the service
life of the pipeline.
The effect of mechanical properties of steel of pipe on pipe system failure is
also of interest. Therefore, two more different values for the yield strength of the
selected pipeline have been studied. Considering the same experimental condition
and 5% acceptable probability of failure, a high strength structural steel water pipe
with nominal yield strength of 475 MPa can operate safely up to 61 years longer
than a steel pipe with 275 MPa tensile strength (Fig. 5.10). From this result, it can
be seen that using a higher grade of steel for water pipes (i.e., higher yield strength)
is generally beneficial; however, this would be more costly in practice.
In order to study the effect of corrosion on the service life of the pipeline, the
sensitivity analysis must be applied to both the corrosion factors used as part of
the nonlinear corrosion model in Eq. (5.18). As can be seen from the results illus-
trated in Figs. 5.11 and 5.12, the multiplication factor a has less impact on pipe-
line failure than its exponential counterpart b. For example, a decrease in a for an
allowed probability of 0.05 increases the time of failure by 52 years in compari-
son with the nominal value (Fig. 5.11). For the same scenario, a 0.5 increase to
the exponential factor b decreases the time to failure by approximately 66 years,
proving that it is highly influential on pipeline failure (Fig. 5.12). It reveals that
corrosion is a significant factor when considering the design of pipelines with
long service lives.
Amongst the four random variables tested, the effect of a and b on the failure of
pipelines is quite remarkable. The difference shown on each of these graphs for
these variables indicates that the sensitivity of failure for studied pipelines is
dependent on the values of these parameters. Accordingly, greater care should be
taken when selecting the values for these variables in future studies.
Overall, this study indicates the importance of random variables on the system
failure of buried steel water pipelines, as well as the considerable impact that the
corrosion process can have over time. Similarly, existing literature also forms
relationships between the probability of pipeline failure and time. Where this
study differs is through applying a time-dependent deterioration method to the
limit states set out in pipeline design manuals. The proposed methodology utilized
the multifailure concept in order to fill the existing gap in similar researches,