Study On Defect Detection And Quantification Of Ultrasonic TOFD

The ultrasonic TOFD (Time of Flight Diffraction, TOFD) is a new non-destructive testing method. It measures and determines size of the defect based on the signal diffracted from tips of the defects. This dissertation focuses on the defect detection and size measurement, the main researches are as following:1. Firstly, based on the TOFD principles, artificial block is designed and manufactured to simulate artificial defects. Type and placement of probes are analyzed and determined. TOFD image is acquired after A-scanning signals are sampled and quantified.2. Secondly, in order to align the lateral echoes which are not well solved by the artificial errors, cross-correlation is introduced to align the A-scanning signals. Lateral echoes, diffraction signals and back wall echo are all well aligned by this method. Considering the noise signal in the imaging during the weld inspection, an adaptive wavelet threshold algorithm is put forward. It selects the optimum threshold in each step scale. Experiments show that this algorithm improves the signal-to-noise radio dramatically while still keeping details of the diffraction signal.In order to improve the accuracy and efficiency of the defect inspection, defect detection algorithm based on the edge detection is proposed. In this algorithm, threshold of edge detection is selected as the local optimum according to the image energy distribution. Results show that this method can identify the defect information in image very well.3. Usually, the maximum relative error in TOFD defect sizing is about 5%. In order to improve its sizing accuracy, post processing is adopted by using the linear method. The maximum relative error of defect sizing is obviously reduced to about 3% after treatment. Finally, the factors affecting TOFD sizing accuracy are analyzed.