Research On Fault Diagnosis Method Of Substation With Redundant Protection

The Northwest750kV network is an important delivery channel of water, fire and windpower and is an extra-high voltage transmission project of the largest scale, the highestvoltage class, the longest and the most complicated line. In the Northwest grid, the750kVsubstation plays an important role, includes750kV,330kV,66kV line, and uses dualprotection configuration. As a result, the topology becomes more complex and the protectionconfiguration is increased. When the fault occurs, the fault information becomes greater andmore complex, so it is necessary for an operator to detect the fault quickly and accurately,locate and recover it. Hence, it is very important to study the fault diagnosis system to ensurethe security and stability operation of the grid.The thesis takes750kV substation as an object to study. The thesis does research byusing fuzzy petri net (FPN), rough sets (RS), information fusion on the uncertainty ofmaloperation and rejection to operate, the imperfection of information souces and the largeamount of alarm information. Result shows the effectiveness of the proposed method.Firstly, as the750kV substation has the characteristics like the uncertainty ofmaloperation and rejection to operate, double net and dual protective configuration, a faultdiagnosis method of substation with redundant protection configuration based on FPN isproposed. In this method, the redundant knowledge of component diagnosis model is studiedand the redundant diagnosis model based on FPN is constructed using information about twoprotective devices. According to the characteristics of double net, the model is decomposedinto two sub-models, the main and redundant model. In which the credibility of the initialinformation is determined by using the joint information entropy and the minimal result setsof fault diagnosis can be obtained through the use of fuzzy reasoning. Result shows theeffectiveness of the proposed method.Secondly, as the750kV substation has the characteristics like the complexity of lines,dual protection configuration, the large amount of alarm information and the imperfection ofinformation souces in the traditional fault diagnosis, a hierarchical diagnosis method ofsubstation with redundant protection configuration based on RS is proposed. In this method,the750kV substation is divided into three regions according to the voltage class and eachregion is divided according to the characteristics of electrical connection. Afterwards, thediagnosis decision table is constructed and simplified by using the information aboutwave-recording devices and the existed fault information, and the minimal reduction can beobtained by using knowledge acquisition method based on RS. Based on this, the comparativesequence and reference sequence is established. Through the use of grey relational analysis,the grey relational grade of attributes and the grey relational credibility of suspicious fault components is determined in the reduction table, and a certain diagnosis result is obtained.Result shows the precision and effectiveness of the proposed method.Finally, the fusion diagnosis model is constructed by using information fusion, in whichthe DS fusion technology is introduced and the results from FPN and RS are integrated in theDS decision-making level. Afterwards, an accurate diagnosis result is obtained through DSdecision rules. Result shows the precision and accuracy of the proposed method.