Study On Insulation Condition Assessment And Fault Diagnosis For Gas Insulated Switchgear

Gas insulated switchgears(GISs) have been widely used in the high voltage, ultra-high voltage transmission system all over the world, since they have the advantages of comapact size, high reliability with little maintenance. At present, so many GISs have been in service with insulation deteriorating increasingly. Thus, it’s necessary and urgent to find an effective way for insulation condition assessment as well as the fault diagnosis method of GIS.By gathering in the related background, specialized standards and regulations, this dissertation establishes a joint condition monitoring platform with the combination of ultra-high frequency detection online in the field and chemical monitoring method offline in the lab. Starting with that, this dissertation studies the index system and the fuzzy evidence theory assessing model of GIS insulation condition assessment as well as the improved evidence theory integrated fault diagnosis approach with the application of fuzzy c-means and improved evidence theory. Some innovative achievements are as follows:To overcome the shortcomings of each individual condition monitoring method, a joint monitoring system is established with the comprehensive utilization of the advatanges of three methods. The experiments of partial discharge under four artificial insulation defects have been conducted and the electrical and chemical characteristics are collected respectively, providing sufficient data for evaluation index extracted.Under the detailed theoretical analysis of the correlation between three kinds of evaluation index with the insulation state, nine electrical index parameter, such as apparent discharge quantity,average time interval of partial discharge and so on, are extracted from UHF PD signals while six chemical index parameters are extracted including concentrations of the main sulfur compound, three concentration ratios c(SO2F2)/c(SOF2), c(CF4)/c(CO2), and c(CF4+CO2)/c(SO2F2+SOF2), as well as the mean-square generation rate of SOF2 and CO2. Also, three characteristical parameters are chosen for preventive tests, including SF6 gas moisture content, gas leakage and the insulation resistance.Since different insulation aspects for GIS are reflected by many factors, whose index always show different evaluation weights, a fuzzy membership model with the combination of the static and dynamic weight is introduced. Further this dissertation proposes a decision-making evaluation model for GIS insulation state with the basic DS evidential reasoning process. The experimental results verify the fuzzy intregrated evidence theory evaluation model, which has higher reliability with more clear and reliable evaluation conclusion.Finally, the fuzzy c-means clustering is introduced for fault diagnosis with individual chmical, electrical information as well as fusion of the two kinds of information respectively. Aiming at problems that full modification of sensor’s evidence information has infirm pertinence and fails to distinguish conflict evidences, a fault diagnosis approach based on improved evidence theory is proposed. Similar evidences and conflict evidences are recognized by means of a conflict evidence criterion to reserve the similar evidences and modify the conflict ones calculating the Jousselme distance between the evidence and the average weight.The dissertation comprehensively analyze the assessing index and model for GIS insulation condition, as well as the intelligent integrated fault diagnosis technology. The research content is the basis of improving the operation safety and reliability for GIS, satisfying the need of power system and equipment’s condition-based maintenance, which also has essential singnificance in theory and practical application to improve the condition assessment of GIS.