Pressure-based leakage characterisation of bulk pipelines

Doctoral Thesis

2019

Permanent link to this Item
Authors
Supervisors
Journal Title
Link to Journal
Journal ISSN
Volume Title
Publisher
Publisher
License
Series
Abstract
Water losses in distribution systems are a huge problem internationally and also in South Africa where more than a third of the water entering the water supply networks is lost through pipe leaks. With water demand increasing due to population growth and urbanisation, water resources are under greater stress and water supply failures are becoming more common. A great deal of work has been done over the past two decades on managing water losses in distribution systems. The Water Loss Task Force of the International Water Association (IWA) played a leading role in this effort, with the “IWA water balance” now widely used as a basis in municipal water loss programs. One of the areas that have received relatively little attention is leakage on bulk pipeline systems. Bulk pipelines connect water treatment plants to bulk reservoirs and distribute water from reservoirs to different towns or water supply zones. Bulk pipes may be operated using pumps or gravity, and generally do not supply consumers directly. It is difficult to determine what the water losses in a bulk pipeline are, as the high flow rates make it impractical or prohibitively expensive to measure flow rates at both ends of these pipelines. Cheaper solutions, such as clamp-on ultrasonic flow meters or reservoir drop tests, are prone to problems and do not have the required accuracy. Due to the lack of reliable and effective methods, water losses on bulk pipes are often assumed to be 2 or 3 %. However, these losses may, in fact, be much greater, and due to the large flow of water transported by bulk pipelines, even small fractions of losses represent large volumes of water. The aim of this project was to develop a method for identifying the size and type of leak present in real bulk water pipelines with minimal disturbance to the operation of the infrastructure. This was achieved by developing a mobile device called the pipe condition assessment equipment (PCAE), which uses pressure testing in combination with the latest models on the behaviour of leaks areas with pressure to assess the condition of the bulk pipeline. To verify the efficacy of the PCAE, the device was first tested on three uPVC pipes with known leakage characteristics in the laboratory (a 12mm round hole, 100mm by 1mm circumferential crack and a 100mm by 1mm longitudinal crack). As expected, the round hole had very small head-area slopes which are negligible, whilst the circumferential crack showed a negative head area slope and the longitudinal crack portrayed a large positive head-area slope. These results were consistent with previous laboratory experiments that investigated the behaviour of round holes and longitudinal and circumferential cracks. Bulk water suppliers and municipalities were then approached to take part in the study. Several bulk pipelines were tested using the PCAE. The results of the field test are discussed in terms of the pre-testing procedures to prepare for the tests, their repeatability and the effectiveness of the device to detect, quantify and characterise leakage on the pipeline. For pipelines with undetectable leakage, a non-intrusive technique that uses a dynamic pressure drop signature from an isolated pipe, to detect and quantify undetectable leakage, was developed. The leakage characteristics of the isolated pipe were estimated from the pressure vs time data. In summary, if the pressure remained constant the pipe was without a leak. If the pressure dropped, a novel mathematical model was fitted to the pressure vs time curve, using the known pipe properties, to determine the characteristics of the leak or leaks present in the pipe. Overall, the PCAE was capable of assessing the extent of leakage on a range of pipe materials, diameters and lengths. It was found that out of the eleven bulk pipelines tested in this study, three could not be tested due to dysfunctional isolation valves and failed connection points. The other eight pipelines that were successfully tested were found to be leaking. The effective initial leak areas for the tested pipelines ranged from 4.88mm2 to 137.66mm2 , whilst the effective head-area slope ranged from 0.0032 mm2 /m to 3.14 mm2 /m and the N1 leakage exponents were found to range from 0.56 up to 1.09. Finally, since there are no well-founded performance indicators for bulk systems, this study also described the findings from analyses of several potential performance indicators using the data from the bulk pipelines tested using the PCAE. The challenges in comparing water losses of different bulk pipelines are highlighted. Based on this, it was found that because every bulk pipeline has its unique characteristic regarding structural parameters such as diameter, pipe material, type of couplings, and operating pressure, the preferred performance indicator for assessing water losses in bulk systems mainly depends on the purpose of the analysis.
Description

Reference:

Collections