On The Signal Detection And Enhancement Technology In Power Cable Fault Diagnosis

The rapid development of cities has made buried cables gradually replace the overhead cable and thus find wide applications.However,with the growth of the social economy,cable load continues to increase as well as brutal ground construction.Thus the buried cables are consequently prone to break down.Moreover,due to reconstructions of power grids,cable relocations and repair,changes in landforms,the original drawings can no longer display the direction and depth of the buried cables,which brings a series of problems to the maintenance and management of buried cables.Because the buried cables span a large space,the searching for fault locations of cable is of much difficulty.Once the faulty cable can not be repaired in time,it will not only waste the manpower and material resources,but also cause power outages.Therefore,the quick and accurate locating of the cable fault points has an urgent social significance.First of all,the fundamentals of cable fault detection,such as fault types,causes and detection methods,are introduced,and the development status of the cable fault detection is also reviewed in this dissertation.Then the key issues of the cable fault detection including the filtering and enhancement processing of the collected signals during the detection process are clarified,and the basic knowledge of signals enhancement and the status of development at home and abroad are further reviewed.Next aimed at the existing challenges in the cable fault detection,this dissertation studys pre-location of the fault,cable path detection,precision location of the fault and signal enhance technology under low SNR and non-stationary background noise in cable fault detection.The main contributions of the dissertation include the following parts:The pre-location of cable faults,i.e.the long-distance cable fault point detection,requires extracting useful weak signals under strong background noise and interference signals.In this dissertation,the waveform characteristics obtained from different cable fault detection methods is studied,and an Empirical Mode Decomposition(EMD)filter based cable fault signal detection technology is proposed.Firstly the EMD is performed on the detected waveform.Then the boundary between the noise and the useful signal in the intrinsic mode component is determined.Next,the wavelet transform is used to further filter the useful signal in the noise dominant component of the signal.By reducing the filtering error to locate through the transmitted waveform and the reflected waveform,we can determine the time difference and thus calculate the fault location.Aiming at the problem that the interference among signals greatly affects the accuracy of detection results under the environment of complex pipeline distribution in cable path detection,the EMD-based fifth-order convergence Independent Component Analysis(ICA)signal extraction technology for cable path detection is proposed.Based on the strong correlation between the signal received and the signal transmitted,and the considerable difference between the signal received and the interference signal and the noise spectrum,the EMD analysis is performed on the signal received.The fifth-order ICA makes the intrinsic mode components decomposed by EMD and the residual signals independent of each other,and automatically extracts the intrinsic mode components whose spectrum has the largest correlation with the spectrum of the known transmitted signal as the final expected signal,thereby achieving the purpose of removing interference and enhancing the useful detection signal.Acoustic and magnetic synchronization method is a common one in the precision location of cable faults.The magnetic signal is difficult to be interfered and easy to be detected,but the amplitude of the sound signal is small and it is easily affected by the surrounding environment noise.In the view of the problem that the background noise of cable faults detection environment is complex and the fault impulse discharge sound signals detected by the cable fault detection fixed-point instrument are completely masked in the noise,this dissertation proposes a sound signal enhancement algorithm based on wavelet packet transform decomposition signal,adaptive filter estimation noise and genetic algorithm optimization reconstruction.Combined with acoustic and magnetic synchronization method,this algorithm is able to calculate the time difference of synchronous transmission between discharge sound signal and magnetic field signal accurately,and improve the precision of fault location.For the fact that the cable fault detection instrument has these disadvantages including that the function is single,the structure is fixed,the corresponding function can not be automatically changed according to the change of the environment,and it can not be selfrepaired when the software and hardware break down.In view of these problems,this dissertation applies the concept of hardware evolution to the cable fault detection to make the cable fault detection develop in the direction of low power consumption,miniaturization,and high reliability.This dissertation uses Field-Programmable Gate Array(FPGA)as the platform of hardware evolution,and proposes the non-permanent elitism Tendency Compact Genetic Algorithm(ne-TCGA)to achieve hardware evolution,which compensates the shortcomings of traditional evolutionary algorithms that occupy large storage space and insufficient search capabilities,improves the efficiency of evolution and lays the foundation for the ultimate realization of multi-function cable fault detection self-evolving system.Based on Empirical Mode Decomposition,the Wavelet Transform,Independent Component Analysis and Optimization Algorithm,the signal enhancement techniques of cable fault detection are studied deeply.Through theoretical analysis,simulation and test application,the problem of extracting useful signals under non-stationary background noise is addressed.At the same time,the signal-to-noise ratio of the cable fault detection signal and fault detection accuracy are improved.Finally,the evolutionary idea is applied to the cable fault detection.And studying a new intelligent cable fault detection system is able to achieve the purpose of detecting and repairing the cable fault points timely and accurately.