Indicator kriging for damage position prediction by the use of electromechanical impedance-based structural health monitoring

Abstract

Some nondestructive techniques of the Structural Health Monitoring (SHM) have improved their analysis in the past decades. Among them, the electromechanical impedance-based SHM technique (EMI-SHM) has been tested in several fields and associated to different statistical methodologies. Considering the nature of the spatial variation of the damage metric data along structures, herein is proposed the use of the indicator kriging method for predicting the existence of a known damage located in the center of an aluminum plate. Maps showing the probability of the damage metric to fall in several value ranges were capable of outlining the areas affected by the damage and predict its location. Comparisons between scenarios with different spacing between PZT patches showed a reduction in the reliability of the model with the increasing of such spacing. Also, for the structure under study, it demonstrates that it is not possible to obtain results by the methodology for distance between sensors/actuators greater than 16.67 cm. However, the results show that this approach can be a viable alternative for using damage metrics to map regions affected by damage and its location.

SGS method applied to damage location and uncertainty modeling for sensor grid in the ISHM

“…The spatial uncertainty of damage metric values was initially evaluated using four regular grids, whose data sets were obtained from Gonçalves et al [14]. In this context, it was considered the following spacings between sensors in the regular grids: (1) 9.09 cm × 9.09 cm (10 × 10 PZTs);…”

“…The spatial correlation between damage metric values was also explored by Gonçalves et al [14]. The indicator kriging (IK) technique was used to estimate the damage's probability of occurrence.…”

Application of Deep Learning Techniques in the Development of Predictive Maintenance and Fault Detection in Electric Motors

“…Several Machine and Deep Learning techniques have been used in conjunction with monitoring techniques in recent years [26][27][28][29]. CNN is a deep learning technique capable of extracting the features and differentiating the data between several conditions, being used for a wide range of image classifications [11].…”

Calibration of a low-cost electromechanical impedance-based structural health monitoring device

“…The electromechanical impedance-based structural health monitoring (SHM) method compares the structure's frequency responses in two states: with and without damage. The property directly related to changes in mass, stiffness, and damping is mechanical impedance, which can be obtained through the electromechanical impedance of piezoelectric elements coupled to mechanical systems (FREITAS et al, 2021;GONÇALVES et al, 2021;PALOMINO, 2008;PALOMINO et al, 2011;PALOMINO et al, 2012;PARK et al, 2003;PARK and INMAN, 2005;REZENDE et al, 2023). This non-destructive evaluation (NDT) technique based on obtaining and comparing variations in frequency response functions uses piezoelectric patches, and the signals related to electromechanical impedance are used to represent the characteristics relevant to the structural health state (SUN et al, 1995).…”