Water Journal : Water Journal March 2011
refereed paper technical features 104 MARCH 2011 water With recent developments in low-frequency accelerometer sensors, there have been vast improvements in leak noise correlation technology for trunk mains from two manufacturers. Sydney Water has tested correlators from four manufacturers. A correlator measures the time delay between the leak noise arriving from one sensor to the next and calculates the leak position based on the pipe material, diameter and length between sensors entered by the operator (Figure 7). Correlation results from leak noise correlations show the pipe length as a two-dimensional scale diagram, with sensor 'A' positioned on the left and sensor 'B' on the right. The positions generating noise appear with peaks above them. To interpret the results, the operator looks for one or two dominant peaks that are at least three times higher than the background level of noise found across the entire length of pipe. Field Testing of Non-invasive Technologies Over the last four years, Sydney Water has performed a range of trials for both the invasive and non-invasive technologies, and this has been documented in a report titled An overview of innovative leak detection technologies for large mains in Sydney, by Aravinda Stanley and Roger Wood. Sydney Water conducted trials with leak noise correlators using both accelerometers and hydrophones on the Bankstown-Ashfield trunk main in 2008, and on the Lansdowne trunk main in 2009. The aim of the project was to test the operational capabilities and sensitivity of the equipment, the threshold levels in terms of leak flow rate, survey distance and leak location accuracy. Equipment from four suppliers was trialled: Echologics, Primayer, Sewerin and Gutermann. After all suppliers reported a high level of background noise at the Bankstown-Ashfield main, a new site was identified. Consequently, the Lansdowne main (600mm CICL), located in bushland at Lansdowne, was selected for the second trial. All the testing was performed on one pipe, which was a 600mm cast iron pipe approximately 40 years old with an average pressure of 50m. A map of its placement is shown in Figure 8. A leak was simulated by tapping in a ball-valve (Figure 9), which was partially opened to simulate small leaks of 3--10L per minute. asset management Figure 8: Map of Sydney Water test site. Figure 9: Leak simulation point. Figure 7: The principle of correlation. Figure 10: Leak simulation showing 5L/M.
Water Journal April 2011