基于虚拟仪器的表面粗糙钢管超声检测信号处理技术研究

发布时间：2018-06-11 17:28

本文选题：钢管 + 超声检测　；参考：《钢铁研究总院》2014年硕士论文

【摘要】：在利用超声波在对钢管进行检测时,其中影响超声检测效果的一个主要因素就是钢管的表面状况。如果钢管材料的表面状况很差,粗糙度高,超声波会在材料的检测表面发生反射、折射作用,这对超声波造成很大的衰减并使得超声回波信号伴随有大量噪声信号,降低了超声检测的灵敏度和信噪比,严重影响后续对超声检测缺陷信号的判别以及定性、定量分析,影响超声探伤结果的准确性。本文结合课题的研究目的和方法,利用现有的实验设备,自行搭建一套完整的超声检测信号处理系统,并在Windows操作系统下,通过美国国家仪器(NI)提供的LabVIEW开发软件设计出了基于虚拟仪器的超声检测信号处理平台。该系统可以激励超声探头发射超声波信号,利用数据采集卡完成超声检测回波信号的模／数转换,并通过USB串口总线把所采集的数据上传到计算机内。在计算机上开发的基于虚拟仪器的超声检测信号处理平台可以对数据采集卡进行状态控制和参数设置,并对所读取的超声检测回波信号做相应的处理工作。利用自行搭建的基于虚拟仪器超声检测信号处理系统,对表面粗糙钢管的超声检测信号进行了相关的处理工作。对表面粗糙钢管的信号处理工作主要包括基于传统傅立叶变换的信号分析处理工作和基于小波变换的信号分析处理工作。基于传统傅立叶变换的信号分析处理工作主要有：①对表面粗糙钢管的超声检测信号进行快速傅立叶变换(FFT),完成表面粗糙钢管超声检测信号的频谱分析；②在频谱分析的基础上,对表面粗糙钢管的超声检测信号做有限长单位冲激响应(FIR)加窗滤波处理,并对所加不同窗函数的滤波效果进行比较分析；③利用短时傅立叶变换(STFT)对表面粗糙钢管的超声检测信号进行时频分析.基于小波变换的信号分析处理工作主要有：①利用小波变换(WT)对表面粗糙钢管的超声检测信号进行时频分析,并与短时傅立叶变换的时频分析结果相比较；②利用小波阈值去噪法对表面粗糙钢管的超声检测信号进行去噪处理,并对软、硬阈值法去噪的结果进行比较分析。研究结果表明,在利用超声检测技术对表面状况达不到标准规定要求的钢管材料进行检测时,如果不进行任何降噪处理,由于检测信噪比较低,很难实现缺陷的可靠检出(工业探伤的信噪比一般应达到6dB～8dB以上)。而通过对检测信号进行适当的滤波或去噪处理后,检测信噪比得到了明显的改善。但是传统的基于傅立叶变换的FIR波技术,在消除噪声的同时会把大量的有用信号也滤除掉,导致波形的失真；而基于小波变换的信号去噪处理,几乎不会造成有用信号的损失,而且可以有效保持缺陷信号的特征形貌,达到更为理想的效果。这为今后表面粗糙钢管的超声检测信号处理技术提供了很好的解决途径。
[Abstract]:The surface condition of the steel tube is one of the main factors that affect the effect of ultrasonic testing. If the surface condition of steel tube is very poor and the roughness is high, the ultrasonic wave will reflect and refraction on the testing surface of the material, which will cause great attenuation to the ultrasonic wave and make the ultrasonic echo signal accompanied by a large number of noise signals. It reduces the sensitivity and signal-to-noise ratio of ultrasonic detection, seriously affects the discrimination, qualitative and quantitative analysis of ultrasonic flaw signals, and affects the accuracy of ultrasonic flaw detection results. Using the existing experimental equipment, a complete ultrasonic detection signal processing system is built by ourselves, and under Windows operating system, The ultrasonic signal processing platform based on virtual instrument is designed by LabVIEW software provided by the National instrument of USA. The system can excite the ultrasonic probe to transmit the ultrasonic signal, use the data acquisition card to complete the A / D conversion of the ultrasonic detection echo signal, and upload the collected data to the computer through the USB serial port bus. The ultrasonic signal processing platform based on virtual instrument developed on the computer can control the state and set the parameters of the data acquisition card. The ultrasonic signal processing system based on virtual instrument is used to process the ultrasonic signal of rough steel pipe. The signal processing of rough surface steel pipe mainly includes the signal analysis and processing based on the traditional Fourier transform and the signal analysis and processing based on the wavelet transform. The main work of signal analysis and processing based on traditional Fourier transform is to perform fast Fourier transform (FFTT) on ultrasonic detection signal of rough surface steel pipe with 1: 1, and to complete the spectrum analysis of ultrasonic detection signal of rough surface steel pipe. (2) on the basis of spectrum analysis, the ultrasonic detection signal of rough surface steel pipe is filtered by finite length unit impulse response (FIR), and the filtering effect of different window function is compared and analyzed. (3) the short time Fourier transform (STFT) was used to analyze the ultrasonic signal of rough steel pipe. The work of signal analysis and processing based on wavelet transform mainly includes: (1) using wavelet transform (WTT) to analyze the ultrasonic detection signal of rough steel pipe, and comparing the results with that of STFT. 2De-noising the ultrasonic signal of rough surface steel pipe by wavelet threshold de-noising method, and comparing and analyzing the results of soft and hard threshold de-noising. When ultrasonic testing technology is used to detect steel tube material whose surface condition is not up to the standard, if no noise reduction is done, the signal-to-noise ratio (SNR) of the detection is low. It is difficult to detect defects reliably. (the SNR of industrial flaw detection should be above 6 dB or more than 8 dB. The signal-to-noise ratio (SNR) of the detected signal is obviously improved after proper filtering or de-noising of the detected signal. However, the traditional Fir technology based on Fourier transform can eliminate the noise and filter out a large number of useful signals, which results in the distortion of the waveform, while the signal de-noising based on wavelet transform can hardly cause the loss of useful signals. Moreover, it can effectively maintain the feature and morphology of the defect signal, and achieve a more ideal effect. This provides a good solution for ultrasonic signal processing technology of surface rough steel pipe in the future.
【学位授予单位】：钢铁研究总院
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TB553;TN911.7

【参考文献】