Date of Award
Fall 2023
Document Type
Open Access Dissertation
Department
Electrical Engineering
First Advisor
Bin Zhang
Abstract
Safe and reliable operation of power plants and power transmission are critical to economy and society. Cables in power generation and transmission are subject to various thermal, chemical, and mechanical stresses, which generally lead to aging and degradation of cable insulation. It is reported that some cables with a projected lifetime of 40 years need to be replaced in 10-15 years. Poorly maintained aged cables can adversely affect power delivery and lead to catastrophic events, such as blackout, fires, and loss of lives. The current research on cable is mainly focused on the detection and localization of hard faults, which include open circuits and short circuits. In most applications, it is too late to take measures when these hard faults occur, especially in those safety-critical electrical systems. The hard faults are gradually formed by insulation degradation under various stresses. If the cable insulation condition can be monitored and the degradation level can be tracked before it fails, some preventative maintenance can be carried out before the system is down. To fill the gap of cable aging continuously monitoring and condition-based maintenance (CBM), this thesis develops a framework of cable insulation degradation fault diagnosis and prognosis (FDP), which combines the joint time-frequency reflectometry (JTFDR) method, inter-digital capacitor (IDC) sensor, frequency modulated continuous wave (FMCW) with signal analysis, feature extraction, Bayesian estimation algorithm for probabilistic cable insulation degradation state monitoring and remaining useful life (RUL) prediction. This thesis first discusses the cable degradation effect according to transmission line theory and the cable aging trend probabilities. A coaxial cable electromagnetic (EM) model is built by multiple S-parameter blocks; the insulation degradation fault is injected into the model by changing the relative permittivity according to the proposed aging models. The JTFDR method is simulated on the cable model with different relative permittivity. The continuous wavelet transform (CWT) based feature extraction method is proposed to generate the health indicators of the insulation. Based on the simulation results, the accelerated aging experiment has been carried out on seven cables with the JTFDR method. The experiment results verify the theoretical analysis and simulation. A cable degradation dynamic model from the simulation result is applied to the experiment data. A particle filter-based Bayesian approach is proposed to estimate the insulation aging level and predict the RUL of cables. The IDC sensors are also applied to measure the capacitance of cable jackets at different aging stages. Based on that, a model is established to describe the dynamics of jacket degradation, and an extended Kalman filter (EKF)-based algorithm for cable jacket degradation FDP is developed. A series of experiments are conducted on multiple cables to verify the IDC sensor and the proposed method. The result shows that the IDC sensors are effective for cable jacket health monitoring, the measured capacitance is suitable for health indicator, and the prognostic results are more accurate. An FMCW setup is built for cable insulation degradation monitoring in both simulation and experiment. The degradation feature is extracted from the FFT spectrum. The EM model is integrated with the accelerated aging experimental FMCW data for PF-based cable state estimation and RUL prediction.
Rights
© 2024, Xuan Wang
Recommended Citation
Wang, X.(2023). Probabilistic Cable Aging Diagnosis And Prognosis With Reflectometry And Capacitance Methods. (Doctoral dissertation). Retrieved from https://scholarcommons.sc.edu/etd/7584