Numerical Study on Urban Infrastructure Diagnosis in Laterally Heterogenous Soils Using Resistivity and Ground Penetrating Radar Techniques

Abstract

Urban environment can be considered a complex system consisting of the engineered pavement physical structure over the buried utilities (water, gas, sewer) network embedded in the background soil environment. Assessment of buried pipeline civil infrastructures using proximal geophysical methods in such instances has to consider possible interferences, difficulties, and incorrect inferences. In this study, we have conducted a numerical modelling investigation to understand and evaluate how electrical resistivity profiling (ERP) and ground penetrating radar (GPR) can be utilised to provide subsurface information that otherwise may not be possible if either one of the techniques is used. A model geometry consisting of a typical pavement structure (asphalt, base/subbase, and background soil) with a single 2 m pipe buried at a depth of 1 m was used. Strong lateral variations in soil type were incorporated over the short pipe section in order to understand the complexities that can arise, especially with ERP measurements. The 3D electrical resistivity measurements were simulated in Comsol using the 4-probe method, while the 2D GPR measurements were simulated in gprMax to obtain the subsurface information. The results from both ERP and GPR were used to develop a practical framework that can be utilised by relevant authorities for proximal condition assessment of their buried assets. It was suggested that ERP can be used as a first level screening tool over the whole pipeline length, followed by discretely selected GPR scans in order to further gain information on the pipe health. This is attractive practically since, following delineations of a large pipe section into shorter subsections, advanced condition assessment approaches that are generally intrusive in nature can then be economically deployed within the subsections suspected of experiencing significant corrosion damage.