基于宽频导波检测锚杆脱粘缺陷的数值仿真

Numerical Simulation of the Defection of Anchor Debonding Defect Based on Broadband Guided Wave

将低频超声导波应用于在役锚杆脱粘的无损检测,难点在于锚杆与周围介质的相互影响,使导波的传播具有复杂性,从而增加了导波信号分析的难度。将选择宽频信号为激励信号,首先对宽频信号在深埋锚杆中的传播特性进行分析,然后将测试信号进行小波包分解,发现不同的缺陷形式对宽频信号中不同频段的响应具有明显的特征。但由于脱粘缺陷和响应信号特征之间并不是单调关系,因此结合BP神经网络技术,建立测试信号与缺陷特征之间的对应关系,来实现对锚杆脱粘缺陷检测。最后,分析不同噪声水平对识别结果的影响。文中方法对于噪声水平小于5%的检测信号,具有较好的识别结果。

The low-frequency ultrasonic guided wave was applied to debonding nondestructive testing of the anchoring bolt in service; the difficulty is that the bolt and the surrounding media will influence each other,so that the guided wave propagation is more complex,as well as increasing the difficulty of the analysis of the guided wave signal. The broadband signal was used as the excitation signal in this article,firstly,it is based on the broadband signal propagation characteristic of the anchor rod in the embedded depth,and the wavelet packet is used to decompose the test signal,it is obvious that the different defects have different frequency bands in the form of the broadband signal response characteristics. Because the defects and the response are not of monotonic relationship among the signal characteristics,therefore,combining wavelet packet decomposition and neural network,the corresponding relation between test signal and damage characteristics is established to realize the defect detection of the anchor. Furthermore,the tested pollution by noise was also used to identify the anchor integrity. The simulation showed an excellent identified results when noise level is less than 5%.