Reconstructing Extended Irregular Anomalies in Pipelines Using Layer-Peeling with Optimization

Abstract

Pipe wall condition assessment is critical for targeted maintenance and failure prevention in water distribution systems. This paper proposes a spatially distributed pipeline condition assessment technique using persistent hydraulic transient waves of a small magnitude (microtransient waves), with a focus on the detection and reconstruction of extended and irregular pipe wall anomalies (e.g., nonuniform blockages and internal or external corrosion that is distributed along a short extent of the pipe). For an extended and irregular anomaly, a pipe’s response to any incident waves will be complex and impose challenges in interpretation. To identify the complex response patterns, an optimization technique has been developed using a differential evolution algorithm to separate the directional impulse response functions (IRFs) and then to differentiate the anomaly-induced response in a directional IRF from noise. A layer-peeling method is then applied to the directional IRF to reconstruct the pipe impedances, which are related to the localized wave speed and pipe wall thickness. Numerical verifications have been conducted on a pipe with a deteriorated section that is assumed to have a constant internal diameter but varying wave speeds along its length (simulating a section with nonuniform external corrosion and wall thinning). The results show that the nonuniformly deteriorated section can be successfully detected and accurately reconstructed using the techniques proposed in this paper.