An unsupervised Learning Algorithm for Fatigue Crack Detection in Waveguides

Ultrasonic guided waves (UGWs) are a useful tool in structural health monitoring (SHM)applications that can benefit from built-in transduction, moderately large inspection ranges, andhigh sensitivity to small flaws. This paper describes an SHM method based on UGWs andoutlier analysis devoted to the detection and quantification of fatigue cracks in structuralwaveguides. The method combines the advantages of UGWs with the outcomes of the discretewavelet transform (DWT) to extract defect-sensitive features aimed at performing a multivariatediagnosis of damage. In particular, the DWT is exploited to generate a set of relevant waveletcoefficients to construct a uni-dimensional or multi-dimensional damage index vector. Thevector is fed to an outlier analysis to detect anomalous structural states. The general frameworkpresented in this paper is applied to the detection of fatigue cracks in a steel beam. The probinghardware consists of a National Instruments PXI platform that controls the generation anddetection of the ultrasonic signals by means of piezoelectric transducers made of lead zirconatetitanate. The effectiveness of the proposed approach to diagnose the presence of defects assmall as a few per cent of the waveguide cross-sectional area is demonstrated