FRP-Substrate bonding quality investigation making use of ultrasonic waves

Abstract

Fiber reinforced polymer (FRP) composite systems are widely used to repair structurally deficient constructions thanks to their good immunity to corrosion, low weight and excellent mechanical properties. The quality of the FRP-substrate interface bond is a crucial parameter affecting the performance of retrofitted structures. In this framework, ultrasonic testing could be used to assess the quality of the bonding [1-2]. In the case of FRP laminates adhesively bonded to roughly inhomogeneous materials, such as concrete, high scattering attenuation occurs due to the presence of heterogeneities. The concrete behaves almost like a perfect absorber generating a considerable number of short-spaced echo peaks that make the defect echo not distinguishable. In order to avoid scattering, waves longer than the discontinuity have to be used, but this expedient makes bonding defects undetectable. The most common practices involve the use of the first echo amplitude, the peak-to-peak or the average amplitude of the signal in a given time window. It is however well known that, when a direct-contact technique is applied, the ultrasonic response in terms of pulses amplitude is affected by two main factors, i.e. the thin film of couplant between the probe and the medium and the pressure of the transducers on the sample. Furthermore the presence of fibers strongly affects the ultrasonic response. The technique developed in this study is based on the energy distribution measurement of ultrasonic signals by means of a statistical parameter, named Equivalent Time Length (ETL), whose expression is the following: ∑N A(tk)2(tk −tAIC)2 ETL= k=AIC+1 , k=[1,2,...,N] (1) Where N is the number of point samples k, A(tk) is the amplitude of the signal at the time tk, tAIC is the onset time of the signal that was selected by using the Akaike Information Criterion (AIC) function [3]. The ETL is sensitive to the presence of bonding defects in the sense that lower values mean higher reflection of wave energy and higher values mean lower reflection and higher penetration through the concrete bonding. In addition to that, it has the advantage not to be affected by the reflected echoes amplitude variations, which often make the amplitude not a reasonable parameter for bonding quality investigation. In order to apply the ETL to the detection of bonding defect, a preliminary numerical study involving a 1-D system with a material discontinuity was performed. A simple situation of reflection and transmission of longitudinal waves incident on the interfaces between three different media was simulated. 2D finite element (FE) analyses were also performed using the commercial software COMSOL Multiphysics 5.0. ∑N A(tk )2 k=AIC+1 The study was conducted experimentally both in vitro, on FRP reinforcements bonded to concrete substrates with imposed well-known defects, and in situ, on reinforced concrete beams in a floor slab and on Seismic retrofitted concrete walls. Laboratory tests were conducted on four concrete specimens with outer dimensions of 180x250x150 mm3, reinforced by carbon fiber reinforced polymer plates set in accordance with the manufacturer's application guideline, using commercial epoxy resin. In order to simulate the lack of bonding of the FRP, three types of defects of well-known dimensions were located on three of the samples by means of the interposition of foils of Teflon between the concrete surface and the adhesive layer (Fig. 1). The samples were named from D0 (absence of defect, i.e. pristine state) to D3 (large defect), according to the size of the employed Teflon foil(s). The setups involved the use of an excitation signal equal to five cycles of sine function generated by a house-built ultrasonic pulser-receiver, two ultrasonic transducers arranged in a pitch-catch mode, and an ultrasonic preamplifier. The experimental study confirmed the numerical findings and may pave the road towards an