考虑衰减补偿的频域反射单元平均恒虚警率电缆缺陷识别方法

Frequency Domain Reflectometry Cell Average Constant False Alarm Rate Cable Defect Identification Method Considering Attenuation Compensation

针对目前频域反射(FDR)法通过人为主观经验判断缺陷,存在误判率高的问题,提出了一种基于单元平均恒虚警率(CA-CFAR)的电缆缺陷识别方法。首先对FDR测试信号在电缆中存在传播衰减,导致缺陷及末端在定位谱中归一化幅值过小而测不准的问题,先使用实测数据求出衰减因子与测试频率的线性关系,再使用在频域解卷的衰减补偿方法,有效削弱了信号在电缆传播过程中衰减过大带来的负面影响;其次通过恒虚警率(CFAR)的动态阈值来提取电缆中缺陷的位置及归一化幅值信息,其中动态阈值根据测试信号幅值的变化动态地选取缺陷判断的阈值,避免了人为主观经验判别容易误判的问题;最后分别在30 m、105 m和500 m实验室电缆以及1500 m在运电缆上对该方法进行测试,测试结果表明采用该方法可以准确地识别电缆全部缺陷,证实了该方法的有效性。

The frequency domain reflectometry(FDR)method using sweep frequency signals is widely used in cable defect monitoring due to its high frequency components in the incident wave and its ability to quickly monitor weak defects.The broadband impedance spectroscopy(BIS)method and reflection coefficient spectrum(RCS)method,developed based on the FDR method,obtain cable characteristic parameters through scanning signal measurement,and can locate local defects in cables with high sensitivity.However,when using FDR to locate defects,spectrum leakage during Fourier transform can result in many"pseudo peaks"on the final localization spectrum.Due to the influence of these"pseudo peaks",it is difficult to accurately identify the position of defect peaks in the localization spectrum,and due to subjective human factors,it is highly likely to lead to misjudgment.At the same time,during on-site testing,the attenuation of the signal will increase with the increase of cable length,resulting in very inconspicuous cable defects during defect detection for some medium and long cables,causing significant economic losses for subsequent repairs.Therefore,the constant false alarm rate(CFAR)method can be introduced for automatic identification of cable defects.It can automatically extract the position and peak value of defects in the positioning spectrum,greatly reducing misjudgments caused by spectrum leakage and subjective human factors,and improving the fault detection rate.Constant False Alarm Rate is a commonly used method in the field of radar target detection.During the process of detecting targets,radar is subject to interference from clutter.The constant false alarm rate detector adaptively changes the detection threshold with changes in the reference unit to achieve CFAR detection of the target.The ultimate goal is to determine whether the object detected by the radar is a target object,which is consistent with our goal of cable defect recognition.What we need is to determine whether the abnormal peak in the positioning spectrum is a defect.Therefore,the CFAR method has a foundation for application.The most basic form of constant false alarm rate is the cell average constant false alarm rate(CA-CFAR),which has a simple algorithm structure and fast calculation speed,and can be effectively applied in cable defect identification.Firstly,in order to reduce the uncertainty of defect measurement caused by signal attenuation effect,a attenuation compensation method is adopted by approximating the attenuation factor with the test frequency and then deconvoluting it in the frequency domain.This effectively weakens the negative impact of excessive signal attenuation during cable propagation.Secondly,in order to reduce the problem of human subjective experience judgment being prone to misjudgment,the dynamic threshold of CFAR is used to extract the location and normalized amplitude information of defects in cables.The dynamic threshold dynamically selects the threshold for defect judgment based on the changes in test signal amplitude,avoiding the shortcomings of human subjective experience judgment.In order to test the effectiveness of this method,tests were conducted on 30 m,105 m,and 500 m cables in the laboratory,as well as on 1500 m running cables.The test results showed that this method can not only accurately extract defect locations,but also effectively solve the problem of defect masking during CFAR detection.