A
ccording to industry analyst firm ARC Advisory
Group, pipelines are relatively easy to maintain.
While pipeline related incidents still present
significant risks to operators and the environment,
more advanced technology is now available to detect pipeline
degradation and leaks. Applying data mining techniques to the
data generated from pipeline maintenance and monitoring
technology, as part of a larger asset management strategy, will
help drive asset enhancement and improve reliability.
Data mining is a combination of data management,
statistical techniques, data analysis and pattern recognition.
From a reliability engineer’s perspective, it may be said
that “the secrets of life are contained in the data,” and the
reliability engineer is challenged to extract these secrets and
use them to inform knowledge based decisions. Asset strategy
decisions are made using a combination of knowledge and
judgment from various sources. However, in today’s pipeline
environment, systems are complex and there is a great deal of
interaction and interdependence between systems. Pipeline
requirements, operating windows, maintenance strategies,
process changes and operating procedures are among the
factors that contribute to varying degrees and rates of
equipment degradation, and eventually lead to a functional
failure.
Developing asset strategies is further complicated by the
fact that some of the knowledge needed to build strategies
resides in subject matter experts (SMEs). Data mining can
As pipelines
age,
maintenance
is the primary
method of restoring performance and
prolonging equipment life.
Paul R.
Casto and Matthew D. Markham,
Meridium, USA,
discuss the data
mining process and its use as a
powerful tool in asset performance
management.
THE DEPTHS
MINING
33
