Extracting Defective Features from Noisy Guided-waves Signals that are Reflected by the Defects Located Inside the Concrete-covered Section of a Pipe

Pipes are widely used in infrastructures for delivering water, chemical and gas. These pipes are partially either covered by concrete or buried in ground. Due to humidity in the atmosphere, they are prone to developing corrosion. However, the early detection of corrosion on the sections that are covered by concrete or soils is nearly impossible. If severe corrosion is allowed to continually occur, it may cause pipe rapture. Consequently, train service will be interrupted and unexpected accidents may occur. To facilitate early detection of corrosion occurring in these partially covered pipes, we investigated the capability of guided waves (GWs) in detecting defects through analyzing the received GWs’ signals that were reflected by various defects. Although GW had found to be effective in locating defects and estimating the extent of corrosion, a proper signal processing tool must be used to extract defects related features from blurred signals due to noise and undesired GWs mode conversion. This paper reports the use of GWs-based method to detect the defects and Continuous Wavelet Transform (CWT) to recover vital defect related features from corrupted GWs signals. To ensure the relevant GW signals can be observed, the signal of emitted tone-burst signal was used as the mother wavelet of the CWT rather than the conventional Morlet wavelet. With the help of such mother wavelet, the CWT can effectively minimize the noise and irrelevant signals. Real in-service pipes were used to verify the effectiveness of such CWT. After its application, the entry and the exit sections of the pipes that are covered by concrete wall can be identified. Most importantly, the defects located within the concretecovered section can also be revealed. the on-site tests prove that the GWs inspection method and the CWT that is using the tone-burst signal as the mother wavelet of CWT can efficiently detect corroded area even they are covered by concrete wall.