Advanced forward modeling and tomographic inversion for leaking water pipes monitoring

Abstract

The timely detection of damage and leakage from pipelines is of extreme importance both environmentally and from a economic perspective. Accordingly, we address the problem of imaging leaking pipes by exploiting single-fold, multi-receiver GPR data. In particular, we devise a tomographic imaging method which, taking advantage of the available knowledge on the investigated 'scenario' (pipe position and size), allows us to detect the presence of a leakage already in its first stages of development. In order to properly design the features of the imaging approach, and test its capabilities in controlled conditions, we make use of synthetic data generated with an advanced full-wave 2.5D Finite-Difference Time-Domain forward modeling solver capable of accurately simulating real world GPR scenarios. Numerical examples obtained in this way show the effectiveness of the developed imaging method.