GNGTS 2015 - Atti del 34° Convegno Nazionale

GNGTS 2015 S essione 3.2 109 pavement layers (wear, binder, base). Regarding this point the demand of end-users is a sub centimeter precision. In order to approach such precision is necessary: a) to increase the accuracy in determining the EM wave velocity in the subsurface that allows to convert in depth data originally acquired as a function of traveltime; b) increase the precision of picking the first sample relative to a actual reflection (first break). The EM wave velocity can be estimated on the basis of the GPR data with different levels of precision (Jol, 2009), but an intrinsic error of the order of 5-10% (Forte et al. , 2014) is expected. For the presented work we used a traditional method for determining the dielectric value of pavement by back-calculating the velocity values by using the drill core data. CASE-1: airport pavements. We tested R.Ex. system on two different sites of an airport pavement in order to: 1) give information of the stratigraphic sequence up to about 2 m of depth from the surface and recognize possible failures or subsidences structures, and lateral variations of the geotechnical properties (Test Site1); 2) verify the thickness of new constructed FRC (Fiber Reinforced Concrete) plates (Test Site 2). The Test Site 1 is an airport transit area where we used the combination of 500 MHz and 1.6 GHz antennas in order to achieve the better tradeoff between resolution and investigation. The depth conversion has been obtained by diffraction hyperbola analysis and by the available drill cores and continuous penetrometric surveys. In Fig. 2a an example of profile is shown. The lateral contrast between the concrete plates (left of profile) and the bituminous aggregate (right of profile) is clear: in yellow we marked the base of concrete and in green the base of the asphalt. The depth of each layer has been stored in a dedicated database implementing a GIS project, which allows to produce depth maps that can give a global overview of the stratigraphic variations of all the imaged layers. In particular, it was observed that the sedimentary contact between granular sub base and clay-silty levels (light blue) deeps below the concrete reaching a depth close to 2.5 m and it disappears in the central sector of longitudinal profiles (map in Fig. 2b). The Test Site 2 is a new aircraft apron where we used the 1.6 GHz and 800 MHz antennas couple. We collected longitudinal transects to cross all the length of the aircraft apron for a maximum length of 160 m. The exact location of already drilled cores on some selected plates was available before geophysical data collection and so we designed the GPR survey crossing all the drill locations for a more constrain data calibration. The GPR analysis evidenced a very strong EMattenuation. This let more difficult and delicate the horizon picking phase aimed at the recognition of the base of the concrete, evidenced by a weak reflector only on the 1.6 GHz data (Fig. 2c). The low reflectivity has been inferred to the high metallic mineral compound within the concrete mixture, as reported by the client. For the depth conversion, we used a variable lateral velocity field, due to the variation of thickness measured by the cores (range from 32.9 to 37 cm) and because of the variability of reflected traveltimes of the horizons. Thus, maintaining the constraints in correspondence of the cores (generally 2/3 cores for each profile) we calculated the gradient of the resulted velocity value between them. The intrinsic error of each picked measure is, from the statistical point of view, equal to the ± the sampling interval. Considering the frequency of 1.6 GHz, the fixed time sampling interval equal to 0.096 ns and the measured velocity values close to 15 cm/ns, we have assumed that the picked horizon lies in a range of ±0.0072 m from the inferred depth. CASE-2 highway survey. 149 Km of recently paved roads were surveyed, recording multi- frequency (800 – 2300 MHz) GPR profiles. We recorded a total of 351 km of data, composed by the acquisition along each road lane in the two track directions. The aim of these surveys was to measure the thickness of the aggregate layers and the depth of the subgrade, imaging and locating any possible lateral variation of thickness and evidencing the main stratigraphic features.