DEVELOPMENT OF A PIPELINE MONITORING SYSTEM FOR DETECTION, LOCALISATION AND CHARACTERISATION OF DAMAGE EVENTS IN PIPES

Abstract

ABSTRACT Mathematical techniques for location of a damage event on a pipe was developed and tested using an experimental test rig. A pulse propagation velocity of 355 m/s was calculated from obtained data when static air was used as the transport fluid and 1538 m/s when flowing water was used as the transport fluid. The reason for the discrepancies between these values and the sound velocity values in air and water was investigated. A 21.2oC and 15.3oC temperature rise above ambient in the pressure pulses were obtained when static air and flowing water used were recorded. A difference of 20mm only was observed between the actual and computed event location when static air was used as the transport medium. When flowing water was used as the transport fluid, a difference of 23 mm was observed. Algorithms for the characterisation of damages in pipes were also developed. These were simulated with the results showing a good agreement between the shapes and magnitudes of the measured original and reconstructed pulses. The simulation was verified with experiments on the test rig. The results showed an underestimation in the magnitudes of the reconstructed pulses in the range of 40 – 45 %. This problem was solved by using a factor K obtained by dividing the maximum amplitude value of the original pressure pulse by that of the reconstructed pulse. A K value of 1.9 was calculated for the particular experimental data set used. Reconstruction of the measured original pulse at a damage location was achieved from combining the measured pulses from two other close locations using the developed Fourier transform based model. A wireless communication device was developed for transmission and processing of measured pressure pulses wirelessly to an analytics platform (ThingSpeak) for real time monitoring. Fifteen experiments were conducted on the experimental test rig using this device at pressure readings of 0.8 bar and 1 bar respectively in the pulse generator. The amplitude of the pulse at sensor 2 denoting the event location using 0.8 bar was 901 mm while the amplitude of the other four pulses were 499, 477, 420 and 346 mm respectively. For the pressure rating of 1.0 bar, the pulse at sensor 2 had an amplitude of 908 mm while the other four sensors were 509, 487, 429 and 355 respectively. In both experimental results, the pulse amplitudes ranged in values according to the distances of the five sensors from the event location with the pulses from the sensors closest to the event location having the highest values and those farthest from the event location having the least values. This showed the effectiveness of the wireless communication device.