Ultrasonic wave propagation in carbon fibre composites is complicated and makes the process of detection and localization of hidden discontinuities challenging. A number of digital and statistical signal processing techniques have been employed to resolve this problem; however the intricate wave propagation phenomena and the imperative need to include the frequency domain data, for digital processing, results in the loss of porosity related information. This paper presents a new technique for non-destructive evaluation (NDE), which stochastically localizes delamination and porosity in CFRP. The technique uses statistical data from reflected signals (A-scans) and twodimensional data sets (C-scans), to develop multimodal probability beliefs.

Related Publications