Field testing on a gas pipeline in service for leak localization using acoustic techniques

Abstract

Acoustic sensing techniques have demonstrated to be promising methods for leak detection. A major challenge lies in the multi modes and dispersion nature in acoustic signals propagating along pipelines, which may cause inaccurate estimation of wave velocity and time delay. To find a best solution for precise leak identification considering a specific application scenario, a field test is conducted on a gas pipeline in service with strong background noise. Various factors affecting the in-situ measurements are investigated, including the ambient noise, the existence of elbow, weld seam, anticorrosive paint or polyethylene coating. The dispersive curve of the test pipeline is theoretically simulated and experimentally measured. Two methods based on discrete wavelet transform and time–frequency spectrum are then employed for leak locating. The positioning accuracy of the two methods are compared and discussed with regards to the influence factors in field testing.