Abstract
On-line fatigue crack evaluation is crucial for ensuring the structural safety and reducing the maintenance costs of safety-critical systems. Among structural health monitoring (SHM), guided wave (GW)-based SHM has been deemed as one of the most promising techniques. However, the traditional damage index-based method and machine learning methods require manual processing and selection of GW features, which depend highly on expert knowledge and are easily affected by complicated uncertainties. Therefore, this paper proposes a fatigue crack evaluation framework with the GW–convolutional neural network (CNN) ensemble and differential wavelet spectrogram. The differential time–frequency spectrogram between the baseline signal and the monitoring signal is processed as the CNN input with the complex Gaussian wavelet transform. Then, an ensemble of CNNs is trained to jointly determine the crack length. Real fatigue tests on complex lap joint structures were carried out to validate the proposed method, in which several structures were tested preliminarily for collecting the training dataset and a new structure was adopted for testing. The root mean square error of the training dataset is 1.4 mm. Besides, the root mean square error of the evaluated crack length in the testing lap joint structure was 1.7 mm, showing the effectiveness of the proposed method.
Highlights
Received: 29 November 2021Structural integrity is a key issue for safety-critical systems such as aircraft, infrastructures, and nuclear plants [1,2]
The guided wave (GW) signal is transformed into a two-dimensional time–frequency spectrogram (TFS) image with the complex Gaussian wavelet transform
The differential TFS between the baseline signal and the monitoring signal is processed as the convolutional neural network (CNN) input in order to amplify the effect of the fatigue crack
Summary
Structural integrity is a key issue for safety-critical systems such as aircraft, infrastructures, and nuclear plants [1,2]. These uncertainties would cause changes to fatigue crack growth itself and GW signals, introducing difficulties for reliable crack evaluation To deal with these problems, machine learning methods are adopted, such as the auto-regression model [16], the Gaussian mixture model [17], artificial neural networks [18], and hidden Markov models [19,20]. Directly adopted the one-dimensional GW signal as the CNN input to localize and evaluate the notched damage in the plate structure. For engineering structures with real fatigue cracks, fatigue crack evaluation is affected by complicated uncertainties, from sources like crack geometries, holes, boundaries, and connections These uncertainties make it difficult to extract GW features for the accurate diagnosis of the crack size, since changes of the GW signals caused by the fatigue crack may be masked.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
