Page 163 - Reliability and Maintainability of In service Pipelines

148 Reliability and Maintainability of In-Service Pipelines

1
0.9 GDD model
0.8
Probability of failure 0.7
Monte Carlo simulation method
0.6
0.5
0.4
0.3
0.2
0.1
0
0 20 40 60 80 100 120 140 160 180 200
Time (year)

Figure 5.18 Verification of the results from GDD model by Monte Carlo simulation
method (external corrosion).

The comparison shows that the probabilities of system failure predicted by
GDD model can be verified by the results of Monte Carlo simulation method, par-
ticularly for small probabilities which are of most practical interest.

5.3.2.1 Sensitivity Analysis

It is known that the failure of a pipe can be affected by different factors, such as
pipe geometry, corrosion coefficients, soil properties, and traffic loads. In view of
the large number of factors that affect the corrosion process and failure modes, it
is of practical significance to identify those factors that affect the failure most so
that more research can focus on those factors. The effect of each variable on the
pipe failure can be estimated by reliability based sensitivity analysis as it was out-
lined in Section 3.5.
To evaluate the sensitivity of the probability of failure to different random
variables, sensitivity indexes are computed for the all 15 random variables.
Figs. 5.19 and 5.20 show relative contribution α 2 and sensitivity ratios (SR) for

25-year time steps, respectively.
It is obvious from the results that the sensitivity indexes of internal pressure
P ðÞ, modulus of elasticity ðE P Þ, deflection coefficient ðK d Þ, impact factor ðI c Þ, sur-
face load coefficient ðC t Þ, wheel load ðFÞ, and pipe effective length ðAÞ are very
low for all values of time.
Among all variables, the relative contributions and sensitivity ratios of the
corrosion parameters (k and n) are highly remarkable. This indicates that corro-
sion is a very important factor for the design of underground pipelines with
long lives.