Sandwich Face Layer Debonding Detection and Size Estimation by Machine-Learning-Based Evaluation of Electromechanical Impedance Measurements

Abstract

The present research proposes a two-step physics- and machine-learning(ML)-based electromechanical impedance (EMI) measurement data evaluation approach for sandwich face layer debonding detection and size estimation in structural health monitoring (SHM) applications. As a case example, a circular aluminum sandwich panel with idealized face layer debonding was used. Both the sensor and debonding were located at the center of the sandwich. Synthetic EMI spectra were generated by a finite-element(FE)-based parameter study, and were used for feature engineering and ML model training and development. Calibration of the real-world EMI measurement data was shown to overcome the FE model simplifications, enabling their evaluation by the found synthetic data-based features and models. The data preprocessing and ML models were validated by unseen real-world EMI measurement data collected in a laboratory environment. The best detection and size estimation performances were found for a One-Class Support Vector Machine and a K-Nearest Neighbor model, respectively, which clearly showed reliable identification of relevant debonding sizes. Furthermore, the approach was shown to be robust against unknown artificial disturbances, and outperformed a previous method for debonding size estimation. The data and code used in this study are provided in their entirety, to enhance comprehensibility, and to encourage future research.

Damage localization using electromechanical impedance technique based on inverse implementation

“…As can be seen, the exponential fit gives better results than the linear fit. This result is in agreement with the physics of the system as has been shown by Cherrier et al 14 and Kralovec et al 15…”

“…Figure 6 shows the DI values for changing distance and angle in the composite structure. For the dependence of the DI on distance, based on the work by Kralovec et al 15 the exponential fit was applied. In addition a linear fit was applied according to the work by Zhu et al 16 The R 2 value for the exponential fit is 0.96 while that for the linear fit is 0.94.…”

“…Very few articles use the EMI technique for damage localization. [14][15][16][17] The techniques for damage localization are inspired by damage mapping techniques used with GWs. Cherrier et al 14 developed the inverse distance weighting technique for damage localization.…”

“…The attenuation factor was developed for unidirectional composite structures as well as isotropic structures. Kralovec et al 15 developed a relationship for the distance with DI (root mean square deviation and mean absolute percentage deviation) using finite element model. They established an exponential relationship and used triangulation for damage detection in an isotropic plate.…”

“…The salient features of the developed technique include, robust damage localization of damage both inside and outside the sensor network and ability to overcome spurious sensor data through the regularization process. In addition, the applicability of the technique over a large frequency band (as opposed to the frequency bands used in the literature [14][15][16] ) brings additional robustness to the technique. Lastly, a fast calibration method requiring only a few measurement runs makes the method resource and time efficient.…”