Detection of multiple structural damages from drive point and cross electro-mechanical impedance signatures. Issue 26 (26th October 2022)

• Record Type: Journal Article
• Authors: Meher, Umakanta; Sunny, Mohammed Rabius
• Is Part Of: Mechanics of advanced materials and structures. Volume 29:Issue 26(2022)
• Page Start: 4738
• Page End: 4758
• Publication Date: 2022-10-26
• Subjects: Structural health monitoring (SHM) -- electro-mechanical impedance (EMI) -- artificial neural network (ANN) -- lead zirconate titanate (PZT) -- root mean square deviation (RMSD) -- covariance coefficient (Cov) -- correlation coefficient (CC); Composite materials -- Mechanical properties -- Periodicals; Composite construction -- Periodicals; 620.118
• DOI: 10.1080/15376494.2021.1937757
• Languages: English
• ISSNs: 1537-6494

Abstract

Abstract: A damage detection scheme using the electro-mechanical impedance based signatures for detection of multiple damages is proposed. A beam and a stiffened plate are considered as two example structures in this study. Cracks of 1 mm width and varying depth have been considered as damages for the beam structure. On the other hand, disbonds between the plate and stiffener along the stiffener length have been considered as damage in the stiffened plate structure. Drive point and cross conductance at two piezoelectric patch transducers in a chosen frequency range have been obtained using the finite element based software ANSYS. Result of ANSYS has been validated through experimental result. Root mean square deviation (RMSD) and the correlation coefficient (CC) of the conductance of the structure from the baseline conductance have been used as damage features. An artificial neural network with one hidden layer has been trained considering the damage features as input and crack depths/disbond lengths at different locations as output. Accuracy of the trained network in detecting crack depths/disbonds has been observed through a set of test cases.