| The content of this thesis is one important part of the major project "Study on the system of Condition Monitoring and Fault Diagnosis&Early-warning for Ventilation and Security of the Power Supply System in Coal Mines", which is the sub-project of "Safety Status Monitoring and Fault Diagnosis Forecast of Mine Ventilation Power Supply System"(No:2007BAK29B05) sponsored by the National Key Technology R&D Program of China during the Eleventh Five-year Plan Period. And it is also one part of the project "The Fault Mechanism and Life Assessment Method of Minng High Voltage Cable Insulation"(No:51377113), which is sponsored by National Nature Science Foundation of China. The content of this thesis is aiming at the research on the aging and failure mechanism of cables under special conditions in the coal mine, especially under combined actions of electrical, thermal, mechanical and environmental stresses.Being the transmission media of electrical energy in coal mine, the mining high voltage cable is the bridge between the power supply and the electric equipment. Its reliability directly affects the continuity of the power supply system and production safety of the coal mine. As is known to all, the characteristic of the PD taken place in cables could reflect the insulation performance. Therefore, it is of great significance to ensure the safety running of cables and to avoid major accident by undertaking the research on monitoring technique and pattern recognition method for PDs of HV mining cables, and grasp the status and defect type characteristics of cable insulation, as well as detecting defects in cable insulation timely and finish pre-maintenance.In this thesis, the10kV cross-linked polyethylene cables and ethylene propylene rubber cables were taken as research objects. Based on the study of domestic and abroad on partial discharge signal detection and pattern recognition in cables, artificial simulation was achieved in insulation defects of the mining high voltage cables, and the test models of electrical tree and cavity in the insulation of cables were designed.(?)-p-q-n spectrums of two defects were obtained to extract the partial discharge characteristics, and a new method for pattern recognition of partial discharge for mining high voltage cables was proposed.The main research content of this thesis is as follows:After study of the literature in domestic and abroad, causes and types of failure for HV cables were conscientiously analyzed. Various methods were researched and compared regarding online monitoring of cables, among which the analysis of PD signal feature extraction methods and pattern recognition methods were focused on.After a brief introduction on basic structures of the mining high voltage cable and the connection mode of underground power supply, causes and types of common failure in mining high voltage cables were analyzed, among which the dielectric loss, the insulation resistance, the grounding current, the partial discharge and the core temperature characteristics, as well as their evaluation standards were focused on.According to defects of mining high voltage cables insulation in the practical operation, two kinds of typical defect models were designed, namely the cable insulation containing cavity and the electrical tree. After analyzing the discharge mechanism of the two kind defects,(?)-q-n spectrums of the two defects were obtained, and the gray image of the partial discharge was constructed.Simulation models were built by using finite element analysis software COMSOL Multiphysics. The electric field distribution of the two cable insulation defects were simulated, and the electric field distribution of electrical tree and the cavity size and shape in the cable insulation were analyzed.The methods on extracting characteristics of the partial discharge were discussed by using the statistical feature extraction method and the fractal feature extraction method. The programs to extract characteristics of partial discharge were designed, and characteristics of partial discharge were taken as the sample data of in PD pattern recognition, which includes fractal dimension, lacunarity, skewness and kurtosis.The extension pattern recognition methods and pattern recognition method based on BP neural network were introduced in details. The pattern recognition program of the partial discharge was compiled by Lab VIEW. The extension method and BP neural network methods were analyzed and compared by simulation of the actual partial discharge data. The experimental results showed that the extension recognition method is of high accuracy, with faster recognition speeds and could be used for pattern recognition of partial discharge in mining high voltage cables, diagnosing defects due to the electrical tree and cavity in cable insulation as well. |
PDF Full Text Request