Page 102 - Reliability and Maintainability of In service Pipelines

Methods for Structural Reliability Analysis 91

effect, the existing literature on reliability analysis and service life prediction of
energy pipelines concrete sewers and ferrous pipes (including cast iron (CI) water
pipes) is presented in this section.
Kleiner and Rajani (2001) defined two main classes of methods for service
life prediction: deterministic and probabilistic methods. Deterministic methods do
not consider variation in any variables that affect pipe behavior and failure, whilst
probabilistic methods consider some or all variables as random variables. In
another classification Clair and Sinha (2012) classified deterioration models avail-
able for predicting service life of water pipes into six categories: deterministic,
statistical, probabilistic, artificial neural networks (ANN), fuzzy logic, and heuris-
tic. The input parameters and output results of pipe deterioration models are
heavily dependent on the type of methodology chosen. To find out the gaps and
limitations of each model, the models are briefly explained here.

3.6.1 DETERMINISTIC MODELS

Deterministic models often use laboratory tests and sample specimens to find the
necessary information, therefore the relationships between components are certain.
Variations and uncertainties in variables are not considered in deterministic methods,
while probabilistic methods consider some or all variables as random variables.
Kaempfer and Berndt (1999) undertook a laboratory experiment to predict ser-
vice life of concrete sewers subjected to sulfide corrosion. They used determin-
istic parameters from an accelerated laboratory test to predict the service life of
concrete sewers. Since their study was in laboratory conditions, they suggested
that for a more accurate result the data from real sewage conditions are necessary.
Rajani et al. (2000) proposed a method to estimate the remaining service life of
CI water pipes by considering that the corrosion pits reduce the structural capacity
of the pipes. The residual capacity of the pipes was calculated by a reiterative
model, based on corrosion pit measurement and the anticipated corrosion rate. The
method was deterministic in the view that it does not consider the uncertainties
involved in all factors contributing to the corrosion and subsequent failures.
A comprehensive deterministic life time assessment has been carried out by
Kienow and Kienow (2001). They performed sulfide corrosion modeling as part
of screening analyses to support the prioritization of sewer evaluation efforts for a
sewer inspection and evaluation in the City of Fresno, CA. The project consisted
of the inspection and evaluation of approximately 90 km of concrete sewers in
sizes ranging from 30 to 70 cm in diameter.
Deb et al. (2002) presented a deterministic model based on analyzing the
growth of corrosion pits on CI pipes, loss of wall thickness, and the strength
reduction of the pipe over time. Kim et al. (2007) developed prediction models