Research On Cable Aging Evaluation Method Combining Low Frequency High Voltage Dielectric Loss And Broadband Low Voltage Dielectric Spectrum

XLPE cable is an important part of power transmission and distribution in the power system,and its aging condition directly affects the safety and stability of the power system.In the actual operation process,the operating environment of power cables is variously affected by various external factors,such as electricity,heat and other factors,which make their insulation performance tested greatly.Therefore,more and more scholars have studied the aging status of cables under different factors.This article introduces a variety of factors that cause different insulation degradation.Among the various influencing factors,thermal aging is one of the most common and common factors that cause its insulation performance to be impaired.Based on the broadband low-voltage dielectric spectrum and ultra-low-frequency high-voltage dielectric loss data of cross-linked polyethylene cables,this paper designs a variety of evaluation schemes through analysis methods such as formula fitting and algorithm classification.The analysis results show that the fusion The electrical spectrum and the characteristics of the ultra-low-frequency and high-voltage dielectric loss data,using the random forest algorithm,can better evaluate the cable aging status,and provide a practical method for practical engineering applications.Based on the insulation and aging principle of XLPE cables,a large number of cable samples and dumbbell samples for accelerated thermal aging are made in this paper.The related equipment is used to test the dielectric parameters and mechanical parameters of the aging samples.Through the analysis of the test curve,it can be seen that the complex real permittivity,imaginary part,and dielectric loss tangent curve of the cable gradually increase with the decrease of the test frequency as a whole,and with the age of the cable.Intensified,the change of the curve in the low frequency band is significantly larger than that in the middle and high frequency bands,indicating that the low frequency band of the test curve has a higher sensitivity to aging,and the test data of the ultra-low-frequency dielectric loss also gradually changes with the increase of the aging time Big.Fit the real and imaginary parts of the complex dielectric constant according to the Cole-Cole model,integrate the low-frequency band of the dielectric loss tangent test curve,and calculate the ultra-low-frequency dielectric loss test data to obtain multiple characteristic parameters.The study found that there is a certain correlation between these characteristic parameters and the aging state.According to the standard and the elongation at break parameters of the dumbbell sample obtained from the test,the aging conditions of the cable samples with different aging time and temperature are classified,and three machine learning algorithms are used to learn and predict the obtained characteristic parameters,and the prediction results are accurate.The calculation of the rate,accuracy,and sensitivity,and evaluation of the effects of different algorithms,the results show that the random forest algorithm that combines the characteristics parameters of broadband low-voltage dielectric spectrum and characteristics of ultra-low-frequency highvoltage dielectric loss has higher accuracy,precision,Sensitivity provides a certain reference for the evaluation of cable aging status.