Audrey Van der Wielen

Nondestructive detection of delaminations in concrete bridge decks

To detect delaminations in concrete bridge decks, nondestructive techniques (NDT) permit a frequent and large inspection of the slabs without damaging structures. This research was devoted to detect simulated defects in twelve repaired concrete slabs. These were scanned with high frequency ground penetrating radar (GPR) with the common offset (CO) and common midpoint (CMP) methods. The electromagnetic waves speed was determined from CMPs. A 3D visualization program was also created to display the CO measurements. The visibility of the inserted defects revealed to be dependent on their lateral extension, their thickness and their constitutive material.

GPR detection of saturated areas into concrete in the presence of a water gradient

In the concrete, saturated areas are most of the time limited by a transition zone, presenting a water gradient. This transition zone can affect the GPR waves reflection and decrease the reflection coefficient by comparison to the coefficient that would be obtained on a sharp interface. To quantify the impact of the water gradient on the reflection coefficient, we performed finite differences simulations. They showed that the reflection coefficient was reduced by 70% if the thickness of the transition zone was larger than 2/5 of the wavelength. Laboratory experiments, using hygrométric sensors for the water content control, confirmed this trend.

Detection of thin layers into concrete with static and CMP measurements

Most concrete bridge decks contain a thin waterproofing layer, whose complex GPR signature can affect the detection of disorders or delaminations into the slab. In this study, we characterized the detection limits of our 2.3 GHz antenna for the detection of thin layers, with static and CMP measurements. In this last configuration, we showed that the radiation pattern and the reflection coefficient estimation are key parameters to use the inversion of the amplitude versus offset (AVO) curves for the estimation of the layer parameters. The theoretical results were compared to the results of FDTD simulations, performed with GprMax2D, and to laboratory measurements, performed on concrete slabs or sand containing thin layers.

Relation between bond quality and impact-echo frequency spectrum

According to EN 1504-10 and ACI Concrete Repair Manual, bond strength and interface quality are the main features of repair system necessary to be assessed. Pull-off test is most commonly used for bond strength evaluation but growing interest in nondestructive techniques (NDT) is recently noted. Impact-echo (IE) is treated as the most promising one for this purpose. The aim of this paper is to analyze an effect of bond quality on stress wave propagation in repair systems. A group of samples has been prepared in order to obtain repair systems of different bond quality. Prior to repair, quality of concrete substrates has been characterized according different techniques: compressive strength, superficial cohesion, surface roughness index and cracking quantification. Than a polymer-modified repair mortar has been applied. After hardening, IE signals have been recorded and pull-off bond strength determined. The relationships between parameters characterizing surface quality, bond strength, IE frequency spectrum and results of wavelet analysis of IE signal have been analyzed. 2 ICPIC 2010 – 13 International Congress on Polymers in Concrete Table 1: Characteristic of tested repair systems Group A Group B Concrete substrate C30 C40 C45 C25 C35 C50 Compressive strength classes C30/37 C40/50 C45/55 C25/30 C35/45 C50/60 Surface preparation PL, SB-D, JH, HD NT, SB-W, SC, LC Sample dimensions 80x60x10 cm 50x50x7cm Repair material PCC (A), Dmax = 2,0mm PCC (B), Dmax = 0,25mm Repair layer thickness 3cm 3cm 2.2 Results of substrate characteristics The quality of substrates was characterized from point of view of their roughness, microcracking and surface tensile strength. The roughness was measured by sand patch test according to EN 1766 resulting Surface Rough Index SRI (Fig.1a). Substrates of Group A can be ranked from polished smooth surface (PL), by dry sandblasted (SB-D) and jack hammered (JH) to very irregular hydrodemolitioned one (HD). In Group B low-pressure waterjetting (LC) has no big influence on profile in comparison to brushed surface (NT), while wet sandblasting (SB-W) and scarification (SC) increase roughness a little. Microcracking of samples of Group A was observed on the cross-section of the 8 cm cores on the near-to-surface layer in the area of 2 cm depth. Density of microcracks was calculated (Fig.1b). It can be concluded, that more aggressive surface preparation technique influence more on microcracking: it was observed two times higher density of microcracks after jack hammering (JH) and hydrodemolition (HD) than after dry sandblasting (SB-D) and polishing (PL). As the aggressiveness of surface treatment of samples of Group B was small, the microcracking was not observed here, although it can be expected a little higher level for scarification. a) 0.0 1.0 2.0 3.0 4.0 5.0 P L SB -D JH H D L C N T SB -W S C surface treatment SR I [m m ] Group A Group B b) Group A 0.00 0.01 0.02 0.03 0.04 0.05 P L SB -D JH H D surface treatment L A [ m m /m m 2 ] Figure 1: Surface Roughness Index, SRI (a) and density of microcracks, LA (b) depending on the method of surface treatment The pull-off test according EN 1542 and ASTM C 1583 04 commonly used for evaluation of bond strength (Fig.2b) was applied for surface tensile strength (fhs) measurement (Fig.2a) including type of failure registration. In case of samples of Group A the concrete quality did not have a major influence on the surface tensile strength after surface treatment as it was for samples of Group B (Fig.3). It can be also observed (Fig.4) that for surfaces jack hammered (Group A) and scarified (Group B), more that 50 % of failures appeared near in the superficial zone (type A1, see Fig.2a). It is probably due to microcracking already mentioned. a) b) A A /B B 15 mm repair material concrete substrate Figure 2: Pull-off for evaluation of surface tensile strength (a) and bond strength (b) A1 A2 A3 15 mm T. Piotrowski, A. Garbacz, A. van der Wiellen, L. Courard, F. Nguyen 3

Detection of near-field, low permittivity layers with ground penetrating radar: Analytical estimation of the reflection coefficient

The reflection coefficient of GPR waves encountering embedded thin layers is commonly estimated using a plane wave, far field approximation. But when the thin layer is situated in the near field of the antenna, the spherical nature of the waves and the possible propagation of a lateral wave into the layer may have a strong influence on the measured reflected amplitude. In this work, we studied through 2D FDTD simulations the behavior of a radar wave interacting with thin layers of different thicknesses. The snapshots and radargrams showed a large influence of the layer thickness on the wave propagation. For the very thin layers, the evanescent wave plays a major role and the plane wave approximation gives a good estimation of the reflection coefficient. For thicker layers, the specific inclination of each multiple reflection has to be taken into account, as well as the lateral wave propagation. On the basis of these observations, we determined which analytical method should be used for the analytical prediction of the reflection coefficient, as a function of the layer thickness.