Research Interests

Unsupervised Learning Algorithm for NDE of Structural Waveguides

Guided Ultrasonic Waves (GUWs) are a useful tool in those structural health monitoring applications that can benefit from built-in transduction, moderately large inspection ranges and high sensitivity to small flaws. The project studies a method based on unsupervised learning algorithms for structural damage detection by means of GUWs. The method combines the advantages of GUW inspection with the outcomes of the Discrete Wavelet Transform (DWT), that is used for extracting robust defect-sensitive features that can be combined to perform a multivariate diagnosis of damage. In particular, the DWT is exploited to de-noise and compress the ultrasonic signals in real-time and generate a set of relevant wavelet coefficients to construct a uni-dimensional or multi-dimensional damage index. The damage index is then fed to an outlier analysis (unsupervised algorithm) to detect anomalous structural states.
The general framework proposed in this study is applied to the detection of crack-like and notch-like defects in seven-wire steel strands, railroad tracks, pipes, and steel girders.

(Top left): Loading configuration for the experiment on the steel beam.
(Top right) Photograph of the part of the tension flange of the steel beam with PZT emitter and a sensor
(Bottom right): results of the outlier analysis. Any datum above the threshold (horizontal line) is associated with damage

Source

Rizzo, P., *Cammarata, M., Dutta, D., Sohn, H., Harries, K.A. (2009). “Unsupervised Learning Algorithm for Fatigue Crack Detection in Waveguides,” Smart Materials and Structures, Vol. 18, 025016 (11pp), doi:10.1088/0964-1726/18/2/025016.

Relevant publications

Dutta, D., Sohn, H., Harries, K., and Rizzo, P. (2009). “A Nonlinear Acoustic Technique for Crack Detection in Metallic Structures,” International Journal of Structural Health Monitoring, 8, No. 3, pp. 251-262.

Rizzo, P., and Lanza di Scalea, F. (2007). “Wavelet-based Unsupervised and Supervised Learning Algorithms for Ultrasonic Structural Monitoring of Waveguides,” Progress in Smart Materials and Structures Research, Ch.8, Ed. Peter L. Reece, NOVA publishers, pp. 227-290. https://www.novapublishers.com/catalog/product_info.php?products_id=4352

Rizzo, P., and Lanza di Scalea, F., (2006). “Discrete Wavelet Transform for Enhancing Defect Detection in Strands by Guided Ultrasonic Waves”, International Journal of Structural Health Monitoring, Vol. 5, No. 3, pp. 297-308.