Study On Location And Assessment Method Of Water Tree Defects In XLPE Cable Based On Reflection Coefficient Spectrum

When the cable is running in wet environment for a long time,water tree aging is very easy to occur inside the insulating media,which not only seriously damages the cable insulation performance,but also may be converted into more destructive electrical tree under the impact of overvoltage.Insulation breakdown may occur in a short time,resulting in power outage.Therefore,it is important to study the effective diagnosis,accurate location and accurate assessment methods for water tree defects to ensure the safe and stable operation of cable lines.To overcome the shortcomings of existing water tree defect diagnosis technologies,such as low sensitivity for local water tree defect diagnosis in long cables,lack of accurate location methods for water tree defect,single assessment feature,subjectivity and limitation of assessment results.In this paper,a method for locating and evaluating water tree defects based on reflectance spectra is proposed,which can accurately locate and reliably assessment local water tree defects in long cables.The main work in this paper is:（1）Based on the transmission line theory,the reflection coefficient spectrum model with multiple water tree defects is constructed.On this basis,the periodicity and attenuation of the reflection coefficient spectrum are analyzed and summarized,which provides theoretical support for the research of water tree diagnosis method based on the reflection coefficient spectrum.The influence of different types of cable defects on distribution parameters and reflectance spectra is summarized,and a defect type identification method based on this feature is presented.The sensitivity of defect type identification is analyzed.（2）A water tree location method based on improved generalized orthogonal method is proposed.The acquisition method of parameters required by the improved algorithm is studied,and the diagnostic sensitivity of the proposed method is analyzed.The influence characteristics of load,water tree aging degree,water tree position,test frequency,cable length,water tree area length,intermediate joint and other factors on the positioning and diagnostic results are clarified.The selection method of the optimal frequency band range for simulation diagnosis is proposed.The results show that the method is effective and the positioning error is less than 0.4%.（3）A water tree defect degree assessment method based on reflection coefficient spectrum is proposed.Four characteristic parameters(kurtosis coefficient K_u,peak value at defect P_def,peak value at cable end P_end and fractal box dimension D)are extracted from reflection coefficient spectrum and diagnosis function of cable.An intelligent assessment model of water tree defect degree based on support vector machine is constructed.The model parameters are optimized by immune genetic algorithm,and the assessment accuracy is as high as 94.57%.