Towards automatic flaw sizing using ultrasonic time-of-flight diffraction

摘要

Despite the recent popularity of ultrasonic time-of-flight diffraction (TOFD) as a reliable non-destructive testing technique for the inspection of weld defects in steel structures, the critical stages of data processing and interpretation are still performed offline manually. This depends heavily on the skill, experience, alertness and consistency of a trained operator. The TOFD data acquisition and display configurations themselves may introduce a host of errors that cannot be accounted for by manual interpretation. This paper presents techniques developed for the automatic sizing and positioning of weld flaws in TOFD data as an essential stage in a comprehensive TOFD inspection and interpretation system to aid the operator by automating the processing and interpretation. Data manipulation and postprocessing techniques have been specifically developed for the automatic detection and sizing of weld defects in TOFD data, significantly reducing the sizing and positioning errors. The system is designed to function autonomously without the need for user intervention, relying instead on information derived from the raw data and the data acquisition parameters. This defect sizing process generates, in near real-time, a quality appraisal of a scanned weld, detailing the location and dimensions of all detected flaws. The results achieved so far have been promising in terms of accuracy, consistency and reliability.