Abstract: For the incomplete fusion defect of the friction welded joint of the piston rod, the ultrasonic non-destructive testing method was used to testing the weld joint of the piston rod and then the ultrasonic echo signal containing noise was extracted. The wavelet noise reduction method was used to filter and reduce the defect signal containing noise.The two-dimensional contour reconstruction of the incomplete fusion defect is carried out by the method of analyzing multi-position and multi-point ultrasonic echo signal. The result shows that: 1) By optimizing the parameters of the wavelet threshold noise reduction algorithm, the signal-to-noise ratio of the ultrasonic echo signal can be improved, and the effectiveness and accuracy of the detection result can be improved. Also it can realize the noise reduction of the defect signal and ensure the integrity of the signal;2) Inspection data was theoretically calculatedto determine the edge points of the defect. Through fitting the edge points of the multi-position defect, the two-dimensional contour reconstruction of the incomplete fusion defect is realized, and the reconstruction result is basically consistent with the destructive inspection result.

Key words: friction welded piston rod, ultrasonic non-destructive testing, wavelet noise reduction, two-dimensional contour reconstruction