Open Access
Subscription or Fee Access
Compressive Sensing and Local Wavenumber Estimations for Fast Damage Imaging with Guided Waves Inspections
Abstract
Many nondestructive evaluations and structural health monitoring techniques rely on the analysis of guided waves propagation in large regions of plate like structures, but such ultrasound inspection techniques are often quite slow and require human interaction. For these reasons, extracting information related to the damage condition of a structure can be a time consuming process. Therefore, it is necessary to have fast and accurate techniques for damage detection and quantification. This research describes a method using compressive sensing (CS) and local wavenumber estimation techniques that can lead to fast scanning and damage detection procedures. The compressed sensing technique reduces the amount of measurements needed, thus achieving faster scanning. In this paper, it is demonstrated that such procedure does not compromise the detection accuracy, as a matter of fact, it allows to improve the performance of damage imaging tasks by removing noise artifacts. In the experiments detailed in this work, guided waves are excited with a piezoelectric transducer bonded to the inspected structure and sensed by an air-coupled probe mounted on a CNC machine for horizontal and vertical scanning. Full wavefields are rapidly reconstructed by applying the compressive sensing technique. Then, local wavenumber domain analysis is performed for processing guided wavefield data. To image the defect itself, we employed the application of both compressive sensing and local wavenumber damage quantification techniques to guided wavefield data for fast damage imaging process. To demonstrate the effectiveness of the proposed techniques, several experiments were performed on aluminum structure, emulating defect with a mass. The results demonstrate that the techniques are very effective in localizing damage with high inspection speed by sampling just the 10% of the Nyquist scanpoints.
DOI
10.12783/shm2017/14040
10.12783/shm2017/14040
Full Text:
PDFRefbacks
- There are currently no refbacks.