Damage Detection and Localization using Random Decrement Technique on Metallic Plates

摘要

Technique with the capability of detecting and localizing damage of structures using naturally operating environments canprovide a possibility of developing more efficient and simpler structural health monitoring systems. This passive sensingtechnique would eliminate the need of active actuation which requires power either from battery or ambients to generatecontrolled excitation source. In a recent study, self-Green’s functions (GF) were reconstructed using auto-correlation (AC),combined with a damage index by comparing the differences in GFs between damaged and pristine metallic panels tolocate the damage. In this paper, random decrement (RD) technique is proposed to reconstruct GF with computationalefficiency. While the RD has been widely used for damage detection and structure parameter extraction in civil structures,in the frequency usually below 1 kHz; this study explores using RD up to 15 kHz for transient wave reconstruction andthen damage localization. The concept is first validated through simulation for a plate structure, and the results show thatthe reconstructed self-Green’s function match well with the one from the auto-correlation technique after approximately10,000 averages of the RD signatures.For experimental verification, a flat aluminum panel and an integral stiffened aluminum panel were subjected to localizedhigh-pressure air from the air compressor. A laser Doppler vibrometer (LDV) was automated to scan a 150 mm × 150 mmarea to create a 13 × 13 2-D array signals for the pristine and damaged structures. GFs were reconstructed using RD andmean square deviation (MSD) was used for creating the damage imaging map.The result shows that the waveforms of GFs from the damaged panel are distinct from those from the pristine panel,provided the scanning positions belong to the same region with the damage. The waveform deviation was observed rathereasily, indicating its damage detection capability. The result also shows that damage imaging results are in agreement withthe real damage locations, which proves damage can be localized via proposed technique.The proposed automation and RD technique reduce inspection time by half in comparison with the one using autocorrelationand manually inspection. It was found that for future structural design, stiffening ribs could be installed in thestructure not only to strengthen the structure but also to assist damage detection with fewer monitoring locations.