No abstract

A sparse-array structural health monitoring (SHM) system based on guided waves was applied to the door of a commercial shipping container. The door comprised a corrugated steel panel approximately 2.4 m by 2.4 m surrounded by a box beam frame and testing was performed in a nonlaboratory environment. A 3-D finite element (FE) model of the corrugations was used to predict transmission coefficients for the A0 and S0 modes across the corrugations as a function of incidence angle. The S0 mode transmission across the corrugations was substantially stronger, and this mode was used in the main test series. A sparse array with 9 transducers was attached to the structure, and signals from the undamaged structure were recorded at periodic intervals over a 3-week period, and the resulting signal database was used for temperature compensation of subsequent signals. Defects in the form of holes whose diameter was increased incrementally from 1 to 10 mm were introduced at 2 different points of the structure, and signals were taken for each condition. Direct analysis of subtracted signals allowed understanding of the defect detection capability of the system. Comparison of signals transmitted between different transducer pairs before and after damage was used to give an initial indication of defect detectability. Signals from all combinations of transducers were then used in imaging algorithms, and good localization of holes with a 5-mm diameter or above was possible within the sparse array, which covered half of the area of the structure.

Sparse-array structural health monitoring systems based on guided waves have been proposed by many authors, current signals being compared with a baseline obtained when the structure was known to be defect free. An image of the structure in the form of a ‘C-scan’ map showing likely defect locations can be produced by combining information from different sensor pairs in the array. It is generally recognized that temperature compensation is essential for the method to work and various compensation methods have been proposed with good results. However, artifacts are commonly seen in the images, making reliable defect location difficult. This is because, as well as the first reflection from the defect that maps to the correct defect location in the image, shadowing effects occur later in the signal and these combine to produce artifacts. This effect can be reduced by appropriate gating of signals, although at a cost in area coverage. If images are formed with multiple different gate locations, the artifact positions and intensities change, but the defect always produces a strong indication. Therefore by combining multiple images, the artifacts can be suppressed and the defect is located more reliably. This strategy has been successfully demonstrated for defects both within the transducer array and close to an edge of both a simple plate and a shipping container door.