PSI - Issue 52

Hongmin Zhu et al. / Procedia Structural Integrity 52 (2024) 679–689 Zhu et al./ Structural Integrity Procedia 00 (2019) 000 – 000

680

2

In the realm of GWSHM, the damage detection and localization are usually implemented based on the baseline subtraction approach (Zhao et al., 2007), where pristine and current measurements are collected and compared from the same monitored structure and environmental condition, e.g., same temperature. Potential damage is then identified if the deviation between the current signals and the healthy baseline exceeds a pre-established threshold. However, such approach is subject to the availability and quality of baseline signals and sensitive to signal distortions induced by varying environmental and operational conditions (EOCs) (Gorgin et al., 2020; Mitra & Gopalakrishnan, 2016), thus being hindered in wider applications. Despite that the temperature compensation (Torres-Arredondo et al., 2014; Yue & Aliabadi, 2020) can be implemented to alleviate the signal distortions, the performance of such operation could much degrade when the temperature deviation between the current and the baseline measurements is greater than a certain range, e.g., 15 ℃ (Mesnil et al., 2023; Yue & Aliabadi, 2020). Temperature compensation techniques also require the reliable computation of correction factors which are challenging to be obtained for large and complex structures (Giannakeas et al., 2023; Yue et al., 2021). These limitations further foster the development of baseline-free techniques in the GWSHM. Recently the baseline-free methods, especially for the sparse transducer arrays, have been gaining more attention in the GWSHM and have been applied to various structures, including but not limited to plates (Sohn et al., 2007), pipelines (Nguyen et al., 2018) and beams (Mustapha et al., 2014). Common techniques in this domain include the time reversal (TR) (Sohn et al., 2007), the reciprocity principle (Huang, Zeng, & Lin, 2018), and the instantaneous baseline (IB) (Anton et al., 2009; Salmanpour et al., 2017) techniques. The main rationale of the TR is the comparison between excitation signals and the reconstructed counterparts while deviations between the signals indicate the presence of damage (H. W. Park et al., 2007). Earlier investigations in this domain mainly focus on the parameter tuning (Anderson et al., 2013; H. W. Park et al., 2007) for a better reconstruction of the excitation signals, while later applications are extended to composite structures (Ciampa & Meo, 2011). Nevertheless, the standard TR requires transducers to re-emit and receive time reversed signals (Liu et al., 2015), which could be impractical in engineering applications. As an alternative, the recently developed virtual time reversal (VTR) (Huang, Zeng, Lin, et al., 2018; Liu et al., 2015; Wang & Shen, 2019) greatly alleviates the hardware issues by replacing the signal re-emission with simple frequency domain operations. The reciprocity-based approach is reliant on the breaking of reversal symmetry between reciprocal paths when a damage exists in a linear medium (Huang, Zeng, & Lin, 2018). This approach is easily implemented and highly flexible such that it can be integrated with other techniques including mode conversion (Zuo & Huthwaite, 2022) and statistical learning algorithms (He et al., 2021). However, the reciprocity method might be susceptible to sensor positionings, especially when the transducers are arranged asymmetrically around the potential damage (Huang, Zeng, Lin, et al., 2018). Comparatively, the IB assumes that the damage information exists in only a few transducer paths (Anton et al., 2009; S. Park et al., 2010) and requires no physical information on guided wave propagation, e.g., time reversibility, reciprocity. Such method distinguishes damaged paths from the similar defect free paths. Main caveat of the IB lies in the sensitivity to transducer configuration (Anton et al., 2009) including the number of transducer paths, the symmetry of the transducer network and the bonding conditions of transducers. In real-life applications, the performance of the baseline-free techniques could be affected by temperature variations and the existence of multiple damages. Sharma et al. (Sharma & Kapuria, 2020) implemented the standard TR on an aluminum plate and found that the time reversibility of Lamb waves is affected by varying temperatures and the excitation frequencies. The influence of the temperatures, however, is insignificant only when the excitation frequency reaches the best-reconstruction accuracy (Sharma & Kapuria, 2020). The damage detection performance of the IB technique is also validated on composite panels under 60 ℃ variations (Mesnil et al., 2023). Meanwhile, the dual damages are considered by (Kannusamy et al., 2022; Wang & Shen, 2019) which deploy the TR-based techniques for damage localisation on metallic panels. Despite these contributions, there is a lack of comprehensive comparisons of the baseline-free techniques regarding damage localisation under varying temperatures as well as for multiple damage scenarios. Moreover, few research have applied baseline-free techniques on anisotropic composite panels considering the potential difficulties posed by the material anisotropy. Corresponding to the limitations mentioned above, this paper aims to compare and analyse the damage localisation performance of few of the baseline-free techniques when being applied to anisotropic composite plates. The remainder part of this paper is organised as follows. Section 2 and section 3 separately introduce the compared baseline-free techniques and the experimental setup for further analyses. Section 4 illustrates the damage localisation