In the field of geophysics and civil engineering applications, ground penetrating radar (GPR) technology has become one of the emerging non-destructive testing (NDT) methods thanks to its ability to perform tests without damaging structures. However, NDT applications, such as concrete rebar assessments, utility network surveys or the precise localization of embedded cylindrical pipes still remain challenging. The inversion of geometric parameters, such as depth and radius of embedded cylindrical pipes, as well as the dielectric parameters of its surrounding material, is of great importance for preventive measures and quality control. Furthermore, the precise localization is mandatory for critical underground utility networks, such as gas, power and water lines. In this context, innovative signal processing techniques associated with GPR are capable of performing physical and geometric characterization tasks. This paper evaluates the performance of a supervised machine learning and ray-based methods on GPR data. Support vector machines (SVM) classification, support vector machine regression (SVR) and ray-based methods are all used to correlate information about the radius and depth of embedded pipes with the velocity of stratified media in various numerical configurations. The approach is based on the hyperbola trace emerging in a set of B-scans, given that the shape of the hyperbola varies greatly with pipe depth and radius as well as with velocity of the medium. According to the ray-based method, an inversion of the wave velocity and pipe radius is performed by applying an appropriate nonlinear least mean squares inversion technique. Feature selection within machine learning models is also implemented on the information chosen from observed hyperbola travel times. Simulated data are obtained by means of the finite-difference time-domain (FDTD) method with the 2D numerical tool GprMax. The study is carried out on mono-static, ground-coupled GPR datasets. The preliminary study showed that the proposed machine learning methods outperforms the ray-based method for estimating radius, depth and velocity. SVR, for instance, calculates depth and radius values with mean absolute relative errors of 0.39% and 6.3%, respectively, with regard to the ground truth. A parametric comparison of the aforementioned methodologies is also included in the performance analysis in terms of relative error.

The presence of aggressive agents (such as chloride ions brought by seawater) in reinforced concrete structures is responsible for the corrosion of the steel rebars. A Structural Health Monitoring technology is developed as a new passive preventive method that would allow for the detection of and for the ability to follow the presence of chloride ions in the cover concrete of reinforced concrete. This technology, referenced as Functional Magnetic Material (FMM), consists on the measurement with an external interrogator of a Magnetic Observable (MO), partially shielded by a patch and corrodible by chloride ions. This paper presents the results of a parametric experimental study, allowing the validation of the concept of this technology, by highlighting the variation of the MO while considering the geometry and the corrosion level of the patch (based on its Relative Mass Loss—RML), as well as the distance between the samples and the interrogator. The results show that the MO of the FMM significantly varies with the increase in the RML of the patch. A 10%-RML for the patch is sufficient for detecting a variation of the MO of the FMM, and the relative variations of the MO are strongly dependent on the distance between the FMM and the magnetometer, as well as the patch’s thickness.

Rebar corrosion represents a major threat to the durability of reinforced concrete structures, primarily in marine environments. Various Non-Destructive Evaluations (NDE) have been developed to detect rebar corrosion; although most of these have delivered successful results, a lack of reliable techniques for proper corrosion prognosis still remains. Under the French Research Agency (ANR) project’s “LabCom OHMIGOD” framework, we introduce here a novel embedded tool to evaluate the level of cover concrete contamination from aggressive agents responsible for causing corrosion. This tool is divided into two parts: a reactive part exposed to corrosion, and a permanent part protected against corrosion. Using magnetic materials in both parts entails “Functional Magnetic Materials” (FMM) and generates a Magnetic Observable (MO). Through the evolution of corrosion on the reactive part, its magnetic properties become affected, which in turn modifies the MO. By means of regular monitoring of MO variations, it is possible to evaluate the aggressive agent ingress. Consequently, by using a variety of FMM tools placed at different concrete depths, it is possible to indirectly evaluate the rebar corrosion risk. This paper presents a numerical model of the tool employing Ansys software. The underlying objective is to investigate tool accuracy through its key parameters, namely, geometry, relative distance to the receiver, coupling effect, and border effect from the rebar. Simulation results demonstrate that by choosing an efficient geometry for the reactive part (25 mm × 25 mm × 1 mm) and position for the tool (between 1 and 3 mm), both a sufficient MO variation range and a negligible coupling effect can be obtained when the FMM is more than 5 cm from any ferromagnetic material.

Among the commonly used non-destructive techniques, the Ground Penetrating Radar (GPR) is one of the most widely adopted today for assessing pavement conditions in France. However, conventional radar systems and their forward processing methods have shown their limitations for the physical and geometrical characterization of very thin layers such as tack coats. However, the use of Machine Learning methods applied to GPR with an inverse approach showed that it was numerically possible to identify the tack coat characteristics despite masking effects due to low timefrequency resolution noted in the raw B-scans. Thus, we propose in this paper to apply the inverse approach based on Machine Learning, already validated in previous works on numerical data, on two experimental cases with different pavement structures. The first case corresponds to a validation on known pavement structures on the Gustave Eiffel University (Nantes, France) with its pavement fatigue carousel and the second case focuses on a new real road in Vend{é}e department (France). In both case studies, the performances of SVM/SVR methods showed the efficiency of supervised learning methods to classify and estimate the emulsion proportioning in the tack coats.