Dalius Misiunas

Leak Location in Single Pipelines Using Transient Reflections

Abstract The use of controlled, small amplitude transient (water hammer) signals for the detection of leaks in pipeline systems is a promising area of research. A pressure transient travels along the system at high speed and is modified by the system during its travel. Leaks within a pipeline partially reflect these pressure signals and allow for the accurate location of a leak by tracing the reflection to its source – a technique commonly known as time domain reflectometry. This paper discusses and provides possible solutions to a number of practical issues associated with leak detection methods of this type, including the impact of the system configuration and methods for detecting leak reflected signals within a transient trace. A set of equations has been derived to locate leaks in a pipeline for all locations of the transient source and measurement stations, and is essential for an automated monitoring system that uses more than one simultaneous measurement of the transient signal. A change detection algorithm is used in this paper to provide an automated approach for detecting leak reflections, which can reduce the ambiguity associated with simple visual inspection of the transient trace. This procedure was validated by both operational and offline (pipeline shutdown) experimental tests conducted at the University of Adelaide.

Burst detection and location in water transmission pipelines

This paper presents results from testing of a burst detection and location technique on a water transmission pipeline. The primary targets of the method are medium and large bursts that are the result of a sudden rupture of a pipe wall or other physical element in the pipeline system. The technique is based on the continuous monitoring of the pressure in the pipeline combined with a hydraulic transient modeling. Analysis of a burst-induced pressure transient wave and its reflections from the pipeline boundaries is used to derive the location and size of the burst. The method has earlier demonstrated promising results on a laboratory pipeline and a dead-end branch of a real water distribution network. Results presented here show that the approach has a potential to be used for burst detection and location in long transmission pipelines. Bursts of different sizes, locations and opening times were successfully detected and located. Different operational regimes of the pipeline were considered. The technique could help to minimize the response time to the pipe failure and therefore reduce the losses associated with a burst and improve reliability of the pipeline operation.