基于灰关联分析的焊缝超声检测缺陷类型识别

Welding defect recognition based on gray correlation analysis in ultrasonic testing

本文提出了一种以灰关联度为测度识别焊缝缺陷类型的方法,并以夹渣、气孔、未焊透、未熔合4种焊缝常见缺陷为对象,对20个实际焊缝样本进行识别计算,并与射线检测结果对比。结果表明,该方法正确识别率达93.7%,具有较好的识别效果。

Based on gray correlation analysis, this paper presents a method to recognize welding defect pattern. Using this method, 20 welding samples with 4 types of welding flaws were calculated and identified, The results were compared with those of X-ray testing. Experiment results show that the method is feasible and effective and the defect recognition accuracy reaches to 93.7%.