Leak Detection in Long Pipelines Systems

Leakage in any fluid distribution network or conveying systems results in
consumption of resources and time, and its impacts affect the on the environment
and the profits for any asset owner. Moreover, a sufficient and applicable leak
detection system, especially, in the oil/gas industry, comes at a high cost and time
consuming, sometimes affects the system’s productivity.
Because of its simplicity and encouraging results from the theoretical, experimental
and real field tests, the water hammer phenomenon promises shows great benefits.
This work has tried to utilise the routine transient events, raising the pump flow
rate, to detect the leak. Also, it attempted to draw on some successful theoretical
techniques, the cross-correlation and its second derivative, to apply on a real field
system. To achieve that, some theoretical and experimental stages had to be carried
out first.
The real system was scaled theoretically to form a laboratory apparatus, so it could
be fitted in a Contaminant Ingress into Distribution Systems (CID) laboratory at the
University of Sheffield. The leak approach was tested by means of a numerical code
for this design before construction of the rig. Then, the experimental rig was
completed and the data collected from it.
In the real field system, the shortage in the data frequency is an obstacle to applying
the approach. The researcher’s colleagues tried their best to improve the data
acquisition system to meet the requirements. Although the improvement made to
the system in terms of the time precision was impressive, the sample frequency
increment was under the desirable level. The signal analysis approach was worked
as expected theoretically, empirically the results were limited. Some trials were
conducted to enhance the signal features. Later, some issues were raised and
clarifications were added.