Risk Assessment of Switchgears in Distribution System Considering Environmental Dependency

Abstract As a link between power system and users, the distribution network plays an important role. Therefore, the power quality and reliability of distribution system are of main significance to both users and power system. The switchgears are one of the most widely used components in the distribution network. Consequently, it is necessary to study the reliability of the distribution switchgears. Traditional researches on risk assessment of switchgears hardly considered the potential threat caused by high temperature, high pressure and humid environment, which is quite common in some areas especially the coastal areas of distribution network. In order to forecast the failure rates of switchgears better, the method for failure rate forecasting considering the environmental factors based on Back Propagation Neural Network (BPNN) and Support Vector Regression (SVR) are advanced in this paper. The related ambient temperature, humidity and the historical failure rate which were obtained from the actual operation distribution data from China Southern Power Grid (CSPG) are normalized firstly. And then create the training function to train the processed data. Next the BPNN and SVR are used to minimize the error between the predicted value and the actual value through iteration. A demonstration based on a coastal city in China is given to prove the validity of the algorithm. The environmental dependency result is compared with the result from the method that only the historical failure rate considered. A general conclusion can be drawn that the SVR method is more accurate than the BPNN method. And it is expected to be widely used in the whole CSPG in the future.