QQ客服 官方微博 反馈留言

题名	超声检测中信号处理研究
作者	简晓明
学位类别	博士
答辩日期	1999-06-18
授予单位	中国科学院声学研究所
授予地点	中国科学院声学研究所
关键词	超声定量检测 自适应滤波 小波变换 神经网络 解卷积 层状介质
中文摘要	制作了包括平底孔、横穿孔和裂纹在内的标准缺陷样品，用中心频率为1.4MHz和1.18MHz的纵波宽带换能器进行了缺陷超声回波信号采集。缺陷回波信号是入射信号与缺陷散射系统响应的线性卷积。用大平底反射信号近似等效入射信号，对缺陷回波信号解卷积，得到缺陷散射系统响应，可以去除或降低其他因素的影响，空出缺陷特征。文中对超声信号解卷积方法进行了分析。应用自适应滤波对缺陷信号解卷积，缺陷特征更加明显。用来判别带有棱边的平面型缺陷与不带有棱边的曲面型缺陷取得了理想的结果。以缺陷回波解卷积得到的缺陷散射系统响应或解卷积为特征。用人工神经网络可以实现缺陷类型识别。对缺陷散射回波和介质底面反射回波幅度，进行一定程度的仪器因素和耦合因素校正后，用人工神经网络可以实现平底孔、横穿孔的类型识别，给出一定范围的平底孔和横穿孔的大小评定。闭合裂纹散射信号非常微弱，应用小波变换和自适应噪声抵消技术可以去除换能器应电背景噪声，提高信噪化，增强裂纹检测能力。同时，还可改善由于应电压造成的检测盲区，提高近表面缺陷的检测能力。论文最后讨论了三层层状介质的二界面检测问题。超声脉冲回波法测量到的回波信号是各界面的脉冲回波序列的叠加。微弱的二界面信号淹没在强的一界面信号之中，直接从回波信号，无法分辨二界面信号，难以进行二界脱粘检测。论文从理论上推导了层状均匀介质的各界面回波信号表达式；对模拟和实测的层状介质回波信号采用自适应噪声抵消方法进行了二界面信号分离处理，实现了二界面脱粘缺陷检测；从多层界面回波信号的叠加关系入手，分析讨论了二界面回波信号特征，理论与实验进行了比较；同时，讨论了耦合因素和仪器因素校正等问题。论文以工业应用为背景，以实验为基础，所得结果有一定实际应用价值。
英文摘要	Standard flaw samples of flat-bottom holes, transverse cylindrical cavities and fractures respectively are made. Two ultrasonic longitudinal wide-band transducers with central frequency of 1.4MHz and 1.18MHz respectively are used to measure flaw scattering echoes. The flaw type cannot be classified directly from flaw echoes and their spectrum. Flaw echo is the convolution of the incident pulse with the flaw scattering system response. The echo from a plane bottom is used to approximate the incident pulse. The flaw scattering system response is obtained by deconvolution of the flaw echo with the echo from plane bottom. By deconvolution to eliminate or reduce the influence of other factors, the flaw features sand out. Many different deconvolution methods are analyzed. The adaptive filtering deconvolution method acquires better results for the wide-band ultrasonic echoes than other methods. The adaptive deconvolution method is applied to real echoes of flat-bottom holes and transverse cylindrical cavities and the flaw scattering system responses of the flaw samples are obtained. Based on the flaw scattering system response, the flaws are classified successfully. The flaw scattering system responses are input to neural networks for automatic flaw type classification. All flaw samples are classified correctly. After the correction of the instrument sensitivity and coupling, the amplitude of the flaw echo and that from flat-bottom are used for flaw type classification and size evaluation through neural networks. The echo from the closed fracture is very weak, and the ratio of signal to noise is very low. Wavelet Transform and Adaptive Noise Canceling are applied to remove or reduce the transducer background noise and enhance the flaw testing sensitivity. As a result, the testing of the flaw near the surface is improved. The ultrasonic bond testing of multilayers are discussed. The ultrasonic echo of multilayers is composed of interface signal components. The second interface signal components are much weaker than the first interface signal components and overlap with the latter. The second interface signal components can not be obtained directly from the multilayers echo. The analytic expression of the echo of multilayers is obtained. Adaptive noise canceling is applied to separate the second interface signal components from the simulated and real ultrasonic echoes successfully. The interpretation for the envelope shape of the interface signal components is made. The theoretical analyses agree well with the real echoes. All the works are based on experiment and real industry application. The research results can be used in real ultrasonic NDT&E.
语种	中文
公开日期	2011-05-07
页码	128
内容类型	学位论文
源URL	[http://159.226.59.140/handle/311008/608]
专题	声学研究所_声学所博硕士学位论文_1981-2009博硕士学位论文
推荐引用方式 GB/T 7714	简晓明. 超声检测中信号处理研究[D]. 中国科学院声学研究所. 中国科学院声学研究所. 1999.

个性服务

查看访问统计

相关权益政策

暂无数据

收藏/分享

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。