Page 326 - Pipeline Pigging Technology

Assessment of pipeline defects

preferential corrosion around the pipe circumference, Fig.3(b). The ability of
the tools to accurately size corrosion on single or repeat runs means that two
types of assessments are possible.

/. Single run: the significance of reported corrosion can be assessed, using
the methods given below. After this assessment, the corrosion can be
categorized, according to the requirements of repair, e.g. Fig.4. However,
where corrosion is still active, the long-term integrity of the line cannot be
easily assessed, and repeat inspections are necessary.

2. Repeat runs: the significance of reported corrosion can be assessed and
corrosion rates estimated. Where corrosion is still active, the long-term
integrity of the line can be evaluated. (Obviously the time between the runs
must be sufficient to allow some measurable corrosion to occur.)

Evaluating corrosion rates

The change in wall thickness readings between two inspections of a
corroding pipeline gives a corrosion rate, Fig.5. This corrosion rate can then
be used to plan future inspections and also to estimate when the pipeline will
need either repair, replacement or downrating.
Fig.5 is obviously a simplification, as an inspection report on a corroded
pipeline may include many thousands of metal-loss readings. Fig.6 gives an
example of the type of wall-thickness data that can be expected.

Application to field data

In a pipeline, each spool can have several hundred metal-loss readings.
Therefore, a variety of wall-thickness measurements are available:

(a) mean metal loss in each spool or the entire pipeline;
(b) maximum metal loss in each spool or the entire pipeline;
(c) distribution of maximum and mean metal loss for the entire pipeline;
(d) distribution of metal loss in a spool.

Following a repeat inspection, changes in all the above will be available.
This causes problems in determining corrosion rates and focussing attention
on spools which may be corroding at a high rate, particularly if the data are
for a long pipeline. It is therefore necessary to somehow 'filter' all the data to
obtain information on the worst spools with the highest corrosion rates.

307