钢包耳轴根部焊缝的相控阵超声检测与信号处理

采用匹配追踪算法和小波变换对低信噪比(2dB和-5dB)钢包耳轴根部焊缝缺陷检测信号进行预处理对比分析,在定制试块上采用64阵元超声相控阵探头对深度为230mm的耳轴根部焊缝进行检测,并且结合基于Gabor原子库的匹配追踪算法对回波信号进行消噪后成像.结果表明,在低信噪比(-5dB)条件下,匹配追踪算法较小波变换有更好的预处理效果;对深度位于220mm和230mm处的钢包耳轴焊缝缺陷的实际检测信号进行去噪处理时,匹配追踪算法能够准确定位缺陷位置并显著提高缺陷处信噪比.

Phased array ultrasonic test and signal process for root flaw of ladle trunnion

The matching pursuit algorithm and wavelet transform were used to denoise and comparatively analyze the simulate signals of ladle trunnion root weld with SNR of 2 dB and --5 dB. In the experiment, a 64 element ultrasonic phased array probe was used to detect the 230 mm depth of trunnion root weld. The ultrasonic echoes of flaw were filtered by matching pursuit algorithm to build the flaw image. The results indicate that the preprocessing denoising effect of matching pursuit algorithm is better than wavelet transform in low signal-to-noise ratio（--5 dB） signals. The matching pursuit algorithm can locate the position of defects accurately and improve the SNR significantly when denoising the actual detection signals which are located at a depth of 220 mm and 230 mm of ladle trunnion weld defect.