PSI - Issue 28

M.Z. Sadeghi et al. / Procedia Structural Integrity 28 (2020) 1590–1600 M.Z. Sadeghi et al./ Structural Integrity Procedia 00 (2020) 000–000

1591

1. Introduction Adhesively bonded joints are gaining more interests in different lightweight design sectors in comparison with other joining techniques due to their advantages like lower structural weight, fabrication cost and improved damage tolerance (Da Silva et al., 2018). Though adhesive joints are widely used, the damage detection and propagation still remain a challenge. Hence, research about various Structural Health Monitoring (SHM) techniques have been increased in the recent years. Various Non Destructive Testing (NDT) methods for damage detection like Ultrasonic inspection (Toyama et al., 2019) and eddy current (Ihn and Chang, 2004) techniques etc., have been developed. But these are limited to single-point measurements, i.e., local measurement of strain at a single point. Therefore, new globalized damage detection techniques which are capable of continuously monitoring damage of the whole structure are developed. Even though SHM uses various NDT techniques, the main difference lies in the operation principle, where NDT methods are applied locally and offline where damage is expected. SHM, on the other hand, provides real time monitoring of the whole structure (Stepinski et al., 2013). The two main groups under SHM are active and passive approaches, which depend on whether an exciter is used during that particular inspection (Tashakori et al., 2018). Amongst the passive methods Acoustic Emission (AE) is the most popular method to identify defect in adhesively-bonded wooden joints (Gozdecki and Smardzewski, 2005). Whereas, recently, amongst the active methods, the Heterodyne method was implemented into SHM for adhesively bonded and mechanically bonded composite plates (Tashakori et al., 2018). Here one of the plates is excited with different frequencies and the transmitted signal is then measured on the adjoining plate. If spikes in the frequency response of the signal is found then it confirms the presence of crack. The other active methods include ultrasonic spectroscopy where information about damage propagation and detection is provided by the resonant frequency and amplitude of response over a wide range of frequencies, sonic vibration technique under which the coin tap test, the mechanical impedance, the membrane resonance and vibro-themography fall into one group. On the other hand, apart from AE many other passive methods are time domain ultrasonics, through transmission, pulse-echo, passive thermography and X-radiography techniques are one group (Guyott et al., 1986). Particular to thermography there are different active techniques like locked Thermography (LT), Pulsed Thermography (PT) and Pulsed Phase Thermography (PPT) and passive techniques like passive Infrared Thermography (IrT) where local increase in temperature on the surface is monitored with infrared camera (Martens and Schröder, 2020). Another method, called the Optical Microscopy, is the one in which a reflected light microscope or sectioning, polishing and microscopically examining of partially damaged specimens is followed. Shenoy et al. (Shenoy et al., 2009) followed the sectioning of partially damaged specimens and examining them whereas Solana et al. (Solana et al., 2007) and Khoramishad et al. (Khoramishad et al., 2010) examined the damaged surface using a video microscope. Other damage measurement techniques include the usage of Chirped Fibre Bragg Grating (CFBG) sensors (Capell et al., 2007) to detect damage in GFRP SLJ under fatigue loading, usage of CFGB sensor and a digital camera by Guo et al. (Guo et al., 2009), combination of ultrasonic and compliance method by Brussat et al. (Brussat and Chiu, 1978). Many works have been devoted into damage detection by using Backface strain (BFS) technique (Graner Solana et al., 2010; Solana et al., 2007; Sugiman et al., 2013; Zhang et al., 1995). In this approach, the strain gauges (SG) are installed on the adherents, on the bonded overlap region, at points where crack initiation is expected to happen. In many works the number and various ways of positioning the SGs like one to three SGs on each surface of the adherents in the bonded region, upto ten SGs on only one face of the adherent, are incorporated (Bernasconi et al., 2014; Crocombe et al., 2002b; Graner Solana et al., 2010; Solana et al., 2007; Sugiman et al., 2013). Majorly, everyone followed one basic rule laid out by (Crocombe et al., 2002a), who's work concentrated on finding the optimal location of the strain gauge on one of the adherent surfaces. Crocombe suggested the positioning of SGs just inside the overlap region (about 0.5mm), subject to easy installation, as it will result in a clear peak of BFS as damage progresses. According to the work done by Solana et al. (Solana et al., 2007), though the author followed the basic rule by Crocombe et al. (Crocombe et al., 2002a), they used six strain gauges, three on each adherent surface in a particular location. By doing so the author also accounted for the advantages of using multiple strain gauges that it was helpful in measuring the strain variation in the width direction and also give an indication of bending effects caused due to improper clamping of the test specimen (if existing). According to the results obtained by Solana et al., the torsional eccentricity of 2.13 0 induced whilst clamping along with relatively lower bending eccentricity of about 0.15 mm increased the initial strain values. Bernasconi et al. (Bernasconi et al., 2014), on the other hand, used ten strain gauges