基于分布式激光超声的带筋构件R区缺陷检测方法

针对带筋构件R区多类型缺陷的快速检测和精确识别等难题,提出基于分布式激光超声的带筋构件R区缺陷检测和分类识别方法。通过建立有限元模型研究了激光超声在带筋构件中的传播规律及其与R区表面裂纹、近表面气孔等缺陷的相互作用机理,进而设计制作了含典型人工缺陷的6061铝合金带筋构件试样,开展激光超声检测实验。实验获得的缺陷时域信号与B-scan特征图像及仿真结果一致,基于反射与衍射原理实现了对带筋构件R区缺陷定位与定量检测,从而验证了所提分布式激光超声检测方法的可行性,为带筋构件制造缺陷的快速检测和分类识别提供了新思路。     

深度学习在超声检测缺陷识别中的应用与发展*

深度学习（Deep Learning）是目前最强大的机器学习算法之一,其中卷积神经网络（Convolutional Neural Network, CNN）模型具有自动学习特征的能力,在图像处理领域较其他深度学习模型有较大的性能优势。本文先简述了深度学习的发展史,然后综述了深度学习在超声检测缺陷识别中的应用与发展,从早期浅层神经网络到现在深度学习的应用现状,并借鉴医学影像识别和射线图像识别领域的方法,分析了卷积神经网络对超声图像缺陷识别的适用性。最后,探讨归纳了目前在超声检测图像识别中使用CNN存在的一些问题,及其主要应对策略的研究方向。     

激光超声检测带过渡圆角平板表面缺陷的数值研究

为了研究激光激发出的超声波在带过渡圆角的金属平板上的传播规律和检测表面缺陷的方法,采用有限元法模拟了该类平板中的激光超声现象,分析了表面波在圆角区域的传播规律和与表面缺陷的作用过程。数值结果表明:激光激发出纵波、横波和Rayleigh波等,其中Rayleigh波主要存在于表面mm量级,并且在过渡圆角处发生模式转换生成了直达波R′和模式转换波R_R等多种表面波;经过过渡区域后的声波在表面缺陷处发生了反射和透射现象,通过B扫图可以检测缺陷的位置。随着缺陷深度的增加,表面波的透射系数不断减小,且透射波R_t和R_st存在0.5 μs左右的到达时间差,该时间差与缺陷深度近似成线性正相关。数值结果为激光超声检测带过渡圆角的平板表面缺陷提供了有价值的参考。     

基于超声相控阵的环焊缝缺陷检测方法研究

针对传统超声探头焦距固定,检测位置的改变就要更换相应焦距的探头而影响检测效率的问题,提出一种基于超声相控阵换能器的环焊缝缺陷检测方法。而超声相控阵具有电子偏转和电子聚焦特性,能在不移动的情况下发射偏转聚焦超声束,有效地解决了上述问题。首先基于超声相控线阵换能器的声场特点,采用数值分析方法,研究了影响声束偏转聚焦性能的几个主要参数。然后给出了与超声相控阵换能器相连接的多通道数据采集系统结构。介绍了单通道声信号的硬件结构及相应的信号处理方法,实现了对换能器中单个阵元的精确延时的控制。实验结果表明,优化设计的超声相控线阵换能器具有较高的检测精度和检测效率。     

激光超声检测中干涉特征提取算法的研究

为了有效地检测结构中的损伤位置及其形状,开展了基于超声导波干涉特征提取算法的损伤成像技术研究。在波场可视化的基础上,从波场历程中进一步分析了损伤对超声导波传播的影响;为了提取损伤附近超声导波干涉的能量,采用时间与空间域上的傅里叶变换对波场中的入射波和反射波进行分离,并根据计算出的干涉能量分布完成了损伤的成像;由于损伤产生的干涉能量在空间上的分布较广,为了减少损伤区域外的干涉能量,提高图像中损伤形状的分辨率,利用实Morlet小波分析的方法对宽频信号进行处理,提取窄带信号分量再进行干涉能量计算,提高了损伤成像的信噪比。     

基于深度神经网络的太阳能电池组件缺陷检测算法研究

针对太阳能电池组件中电池片出现隐裂导致整片电池破碎,最终影响整个组件发电量的问题,在对电池组件光致发光（PL）图像待检测区域筛选定位的基础上,提出了一种利用卷积神经网络（CNN）进行电池组件隐裂缺陷检测的方法。首先利用PL成像方法获取电池组件图像,然后对图像进行预处理,基于聚类的方法对待检测目标区域进行筛选定位,最后利用3种不同结构的卷积神经网络模型对电池片进行缺陷检测,并进行准确率对比,使最优识别准确率达到99.25%。实验结果验证了该方法能准确地检测出太阳能电池组件的隐裂缺陷。     

卷积神经网络的缺陷类型识别分析

该文提出一种基于卷积神经网络直接对阵列超声检测原始信号进行缺陷类型识别的方法,该方法无需对超声回波原始信号进行特征提取.文章研究对比了不同卷积神经网络及其优化的识别性能.首先采用超声相控阵系统对不同试块上的平底孔、球底孔、通孔三种缺陷进行超声检测,然后利用LeNet5、VGG16和ResNet三种卷积神经网络对一维和二...     

多层结构中脱粘缺陷的超声检测方法

固体火箭发动机多层装药结构中的脱粘类型、位置和尺寸决定了其对整体安全性能构成的威胁程度。该文研究了多层结构中脱粘缺陷的超声检测方法,通过对不同脱粘缺陷超声脉冲回波的特征分析与统计,实现缺陷的定性、定位和定量。首先,采集含有多类脱粘缺陷的粘接结构的超声脉冲回波信号,分析信号中主能量波包所代表的声程,提取五种声程的波峰时刻和幅值作为特征值,组建已知脱粘类型训练样本并输入至BP神经网络,实现特征值域到类别域的非线性映射,即脱粘类型分类;其次,采用阈值法确定缺陷的界面位置;最后,提出分段线性插值-相关性定量法将待检测缺陷的定量结果缩小到±2 mm以内。该文利用COMSOL有限元仿真和实验操作验证了多层粘接结构中脱粘缺陷的定性、定位和定量方法的可行性和可靠性。     

基于卷积神经网络的气体绝缘组合开关盆式 绝缘子螺栓松动检测方法*

盆式绝缘子是GIS的关键绝缘器件,它与两侧气室法兰通过螺栓进行紧固连接,当螺栓松动时会导致盆式绝缘子应力分布不均,严重时会引起绝缘子破裂,从而影响GIS运行的安全性和可靠性。文章搭建了盆式绝缘子螺栓松动超声波检测系统,以获取不同螺栓不同工况下的超声信号,基于卷积神经网络对超声信号进行特征提取,并且与BP神经网络的训练结果进行对比分析。实验结果表明,卷积神经网络可以自动提取GIS盆式绝缘子螺栓松动特征量,对十种螺栓松动工况的识别准确率达到100%,相比于BP神经网络具有较高的识别准确率,该方法可以直接用于盆式绝缘子螺栓松动检测。     

基于分布式超声无损检测方法的零件内部缺陷检测

超声检测广泛应用于工业检测,比如超声相控阵检测法和超声A扫应用于零件内部缺陷检测。然而,这些方法可以检测出缺陷位置却很难精确地检测缺陷尺寸,缺陷定量成了急需解决并且很有意义的问题。本文提出了一种分布式超声无损检测方法,将超声探头均匀布置在检测表面,每一个超声探头可以同时发射和接收超声信号,通过对接收到的信号进行处理来重构缺陷轮廓。基于分布式超声无损检测方法,重构零件的人造缺陷并建立相应的声学仿真模型。通过多项式拟合法和聚类法分别处理实验和仿真所获得的数据并重构缺陷轮廓。实验结果和仿真结果显示重构的椭圆形缺陷和正方形缺陷具有一定的精度。结果表明分布式超声无损检测方法有潜在的应用价值和理论意义。     

Laser ultrasonic propagation imaging method in the frequency domain based on wavelet transformation

A wavelet-transformed ultrasonic propagation imaging method capable of ultrasonic propagation imaging in the frequency domain was developed and applied as a new structural damage or flaw visualization algorithm. Since the wavelet-transformed ultrasonic propagation imaging method has strong frequency selectivity, it can visualize the propagation of ultrasonic waves of a specific frequency (for example, to isolate ultrasonic mode of interest and a damage-related ultrasonic wave). The strong frequency selectivity of the wavelet-transformed ultrasonic propagation imaging method was demonstrated, isolating only the zeroth-order asymmetrical mode of the fundamental Lamb wave modes in an anisotropic carbon fiber-reinforced plastic plate with a thickness of 5 mm. The wavelet-transformed ultrasonic propagation imaging method can also convert a complex time domain multiple wavefield into a simple frequency domain single wavefield. This feature enables easy interpretation of the results, and facilitates the precise evaluation of the location and size of structural damage or flaws. We demonstrated this capability by detecting a disbond in a sandwich structure made of Al-alloy skins and a foam core. A disbond with a diameter of 20 mm, which is representative of a common manufacturing flaw, was successfully detected, localized, and evaluated. Since a method to determine the allowable maximum pulse repetition frequency depending on target materials and structures was found by investigating the residual wave caused from the previous laser impinging, our laser ultrasonic system can scan rapidly the target with an optimal pulse repetition rate. In addition, the proposed wavelet-transformed ultrasonic propagation imaging method can visualize damage or flaw without the need for reference data from the intact state of the structure. Hence, we propose the wavelet-transformed ultrasonic propagation imaging approach for automatic inspection of in-service engineering structures, or in-process quality inspection in manufacturing.     

The effect of surface cleanness of thin defects on the reflection of ultrasound

In a previous paper, the authors described experiments in which simulated defects of thickness 20–200 nm were detected ultrasonically. The defects were produced by creating thin gaps between the surfaces of optical flats. A relation between reflectivity and cleanness of the surfaces was noted.This paper describes new experiments and some theoretical considerations which further explore the relation between cleanness and reflectivity. It is concluded that the surface properties of the thin defects tested affect the reflectivity. In particular, it appears that the surface energy of the defect attracts additional material, which provides a mechanism for transmission of sound across the defect.     

微小层片型缺陷的超声非线性区域检测技术*

针对微小层片型缺陷的超声检测效率低、检测能力弱、易受主客观因素干扰的问题,提出弹簧扁钢中微小层片型缺陷的非线性超声区域检测技术。首先,优化探头夹持装置并提出区域检测方法,提取稳定可靠的检测信号;其次,提出相对非线性系数均值及波动系数表征缺陷分布;最后,基于斯皮尔曼次序相关系数分析扁钢各类缺陷与非线性超声系数均值的相关性。研究结果显示:两种非线性超声特征参数中的波动系数具有更高的缺陷敏感度,可表征扁钢中缺陷分布;波动系数与扁钢中微小层片型缺陷的相关系数比与体积型缺陷的相关性系数大得多,特别适用于微小层片型缺陷的检测。     

Using Convolutional Neural Network and a Single Heartbeat for ECG Biometric Recognition

The electrocardiogram (ECG) signal has become a popular biometric modality due to characteristics that make it suitable for developing reliable authentication systems. However, the long segment of signal required for recognition is still one of the limitations of existing ECG biometric recognition methods and affects its acceptability as a biometric modality. This paper investigates how a short segment of an ECG signal can be effectively used for biometric recognition, using deep-learning techniques. A small convolutional neural network (CNN) is designed to achieve better generalization capability by entropy enhancement of a short segment of a heartbeat signal. Additionally, it investigates how various blind and feature-dependent segments with different lengths affect the performance of the recognition system. Experiments were carried out on two databases for performance evaluation that included single and multisession records. In addition, a comparison was made between the performance of the proposed classifier and four well-known CNN models: GoogLeNet, ResNet, MobileNet and EfficientNet. Using a time–frequency domain representation of a short segment of an ECG signal around the R-peak, the proposed model achieved an accuracy of 99.90% for PTB, 98.20% for the ECG-ID mixed-session, and 94.18% for ECG-ID multisession datasets. Using the preprinted ResNet, we obtained 97.28% accuracy for 0.5-second segments around the R-peaks for ECG-ID multisession datasets, outperforming existing methods. It was found that the time–frequency domain representation of a short segment of an ECG signal can be feasible for biometric recognition by achieving better accuracy and acceptability of this modality.     

Automatic ECG Classification Using Continuous Wavelet Transform and Convolutional Neural Network

Early detection of arrhythmia and effective treatment can prevent deaths caused by cardiovascular disease (CVD). In clinical practice, the diagnosis is made by checking the electrocardiogram (ECG) beat-by-beat, but this is usually time-consuming and laborious. In the paper, we propose an automatic ECG classification method based on Continuous Wavelet Transform (CWT) and Convolutional Neural Network (CNN). CWT is used to decompose ECG signals to obtain different time-frequency components, and CNN is used to extract features from the 2D-scalogram composed of the above time-frequency components. Considering the surrounding R peak interval (also called RR interval) is also useful for the diagnosis of arrhythmia, four RR interval features are extracted and combined with the CNN features to input into a fully connected layer for ECG classification. By testing in the MIT-BIH arrhythmia database, our method achieves an overall performance of 70.75%, 67.47%, 68.76%, and 98.74% for positive predictive value, sensitivity, F1-score, and accuracy, respectively. Compared with existing methods, the overall F1-score of our method is increased by 4.75~16.85%. Because our method is simple and highly accurate, it can potentially be used as a clinical auxiliary diagnostic tool.     

基于模态识别与分离的铝板激光超声检测

用Nd:YAG激光器在2 mm厚的铝板上激发激光超声,通过激光反射镜偏转进行点阵扫描激励,用声发射传感器接收在铝板中传播的激光超声信号,通过信号放大和数据采集得到激光超声在铝板中传播的数据。对上述数据进行二维傅里叶变换得到了铝板中激光超声的模态分布和频率分布,接收到的激光超声是S0和A0两个模态多种频率的Lamb波。利用两个相同的探头对称地压紧在板的两个表面上,对两个探头接收到的信号进行相加和相减,可以将S0和A0模态分别提取出来。对单一模态的Lamb波再进行滤波,得到了窄频带单一模态的Lamb波。采用该方法实现了从杂乱的激光超声原始信号中提取出清晰的伤信号,并且得到了不同模态在不同频率范围的检测结果。     

超声波透射法检测水泥模块的缺陷

介绍了超声波透射法检测水泥模块中缺陷的基本原理,并通过对自制水泥试块的检测,研究了不同缺陷对声学参量的影响.测试结果表明:有缺陷模块会使检测超声波比正常传播时声时变大(声速变小),波幅明显衰减,波形产生畸变,频率也会发生异常波动.因此,评价水泥模块的缺陷时要将声时、幅值、频率、波形结合起来,综合分析异常变化来确定缺陷的性质.