The design of a novel mother wavelet that is tailor-made for continuous wavelet transform in extracting defect-related features from reflected guided wave signals

Abstract

Guided waves are increasingly being researched and applied for use in defect detection. Because different guided wave mode conversions and noise are inevitably generated by wall-covered sections of pipe, the received signals that contain the defect information are often overwhelmed by these undesired disturbances. Hence, the received signals must be analysed by an effective signal processing method so that the critical defect information can be extracted from the blurred received signals. Researchers have used different signal processing methods, such as continuous wavelet transform (CWT), to filter out the unwanted part of the received signals. Different filter bands can be generated from the mother wavelet of the CWT. The selection of a suitable mother wavelet can minimise the levels of noise and unwanted modes. However, the selection of the mother wavelet is an intensive trial and error process. In this paper, a new approach for selecting the most suitable mother wavelet for the inspection process was developed. Instead of using the Morlet wavelet as the mother wavelet, as suggested by other researchers, here propose the use of the emitted tone-burst signal as the mother wavelet for generating different filter bands of CWT. This type of novel mother wavelet is called the tone-burst wavelet. To verify its effectiveness, tone-burst wavelet was tested via simulation and in field tests. Simulated models of pipes were built using the finite element analysis tool to investigate the propagation characteristics of guided waves along the inspected pipes, especially on the wall-covered section. After the successful completion of the simulation, laboratory tests were conducted using building pipes that had corroded areas in the wall-covered sections. Finally, field tests were conducted using in-service building pipes erected in a tall building. The proposed tone-burst wavelet de-noising method was found to be effective in extracting the defect-related signals. Hence, the location and severity of the corrosion were successfully determined. Moreover, the results showed that the tone-burst wavelet was superior to the conventional Morlet wavelet when used as the mother wavelet of CWT. With the help of the proposed method, the likelihood of pipe ruptures that can trigger gas leakage and explosions can be minimised.