相控阵超声检测缺陷识别与分类研究进展  

Research progress on defect recognition and classification by using phased array ultrasonic testin

作　　者：刘春华[1] 周长霖 陈晓辉[1] 陈钦[1,2] LIU Chunhua;ZHOU Changlin;CHEN Xiaohui;CHEN Qin(Facility Design and Instrumentation Institute,China Aerodynamics Research and Development Center,Mianyang 621000,China;State Key Laboratory of Aerodynamics,China Aerodynamics Research and Development Center,Mianyang 621000,China)

机构地区：[1]中国空气动力研究与发展中心设备设计与测试技术研究所,绵阳621000 [2]中国空气动力研究与发展中心空气动力学国家重点实验室,绵阳621000

出　　处：《无损检测》2023年第12期31-37,共7页Nondestructive Testing

摘　　要：相控阵超声技术是近年来无损检测领域的重点研究方向之一,已经取得了飞速发展,其中基于相控阵超声成像的缺陷识别与分类是研究的热点之一。概述了相控阵超声无损检测的基本原理,介绍了具有代表性的缺陷识别与分类算法,包括支持向量机、人工神经网络、遗传算法、神经进化算法和基于深度学习的算法。最后指出了现有缺陷识别与分类算法面临的挑战,并结合实际提出了相控阵超声缺陷识别与分类的发展方向。Phased array ultrasonic technology is one of the key research directions in the field of nondestructive testing in recent years and has made rapid development,among which defect recognition and classification based on ultrasonic phased array imaging is one of the research hotspots.This paper summarized the basic principles of ultrasonic phased array nondestructive testing and introduced representative defect recognition and classification algorithms,including support vector machines,artificial neural networks,genetic algorithms,neural evolutionary algorithms,and algorithms based on deep learning.Finally,it pointed out the challenges of existing defect recognition and classification algorithms and put forward the development direction of ultrasonic phased array defect recognition and classification.

关 键 词：相控阵超声 超声无损检测 超声成像 缺陷识别与分类 

分 类 号：TG115.28[金属学及工艺—物理冶金]