Page 70 - 无损检测2024年第六期

缺陷的智能识别与分类专题
缺陷的智能识别与分类专题

DOI：10.11973/wsjc202406007

基于 SHAP 可解释性的焊缝缺陷类型超声识别

XGBoost 模型

陈明良 ，马志远 ，张东辉 ，付冬欣 ，廖静瑜 ，林 莉 1
1
2
2
1
2
（1. 大连理工大学 无损检测研究所，大连 116024；2. 中国核工业二三建设有限公司，北京 101300）
摘 要：针对焊缝缺陷机器学习超声识别过程中存在特征冗余、可解释性差等问题，提出了一
种基于SHAP可解释性的焊缝缺陷超声识别XGBoost（极限梯度提升）模型。在碳钢焊缝试样上加
工4类典型缺陷，采用横波斜入射法采集超声反射回波信号，分别提取16个时域特征、16个频域特
征以及3个信息熵特征。计算SHAP值并选择其前8个高贡献特征构建特征子集，利用交叉验证和
网格搜索优化XGBoost模型进行缺陷识别。试验结果表明，4种缺陷识别的平均准确率为96. 7%；
其中，横通孔的识别效果最佳，精确率、召回率和F 1-score 均达到100%，三角槽次之，方形槽略差，矩
形槽的识别结果最差，其精确率、召回率和F 1-score 均为93. 3%。最后，讨论了高贡献特征与缺陷类
别之间的相关性，并对特征贡献差异及其原因进行了分析。
关键词：超声检测；缺陷分类；XGBoost模型；特征选择；SHAP
中图分类号：TG115.28；TB553 文献标志码：A 文章编号：1000-6656(2024)06-0036-07

XGBoost model for ultrasonic recognition of weld defect types based on SHAP interpretability

CHEN Mingliang , MA Zhiyuan , ZHANG Donghui , FU Dongxin , LIAO Jingyu , LIN Li 1
2
2
1
2
1
（1. NDT & E Laboratory, Dalian University of Technology, Dalian 116024, China;
2. China Nuclear Industry 23 Construction Co., Ltd., Beijing 101300, China）
Abstract: Aiming at the problems of feature redundancy and poor interpretability in the process of ultrasonic
recognition of weld defects based on machine learning, an XGBoost model for ultrasonic recognition of weld defects was
proposed based on SHAP interpretability. Four kinds of typical defects were machined on carbon steel weld samples. The
ultrasonic reflection echo signal was collected by shear wave oblique incidence method, and 16 time-domain features, 16
frequency-domain features and 3 information entropy features were extracted respectively. SHAP values were calculated
and 8 high-contribution features were selected to build a feature subset. Cross-validation and grid search were utilized to
optimize the XGBoost model. The feature subset was used as input to identify the defect types. The results showed that
average recognition accuracy of the four defects was 96.7%. Among them, the recognition effect of the transverse hole was
the best, and the precision, recall and F all reached 100 %, followed by the triangular groove, and the square groove
1-score
was slightly worse. The recognition result of the rectangular groove was relatively poor, and its precision, recall and F
1-score
were all 93.3%. Finally, the correlation between high contribution features and defect categories was discussed, and the
difference of feature contribution and its causes were analyzed.
Key words: ultrasonic testing; classification of defect; XGBoost model; feature selection; SHAP

收稿日期：2024-01-29
超声检测具有灵敏度高、操作简单、适用范围广
作者简介：陈明良（1999—），男，硕士研究生，主要研究方向为超
[1]
等优点 ，是金属材料焊接质量控制的重要手段，根
声检测缺陷识别
据缺陷回波特征识别焊缝缺陷类型是质量评定的重
通信作者：林 莉（1970—），女，教授，博士生导师，主要研究方
要环节 。随着人工智能和信号处理技术的发展，缺
[2]
向为超声检测与评价，linli@dlut.edu.cn
36
2024 年 第 46 卷 第 6 期
无损检测