Application of ARMA Model in Prediction of Development Trend of Partial Discharge

: Nowadays, gas insulated switchgear (GIS) has been widely used in power systems. Due to some external factors, there may be defects in GIS, and some minor defects are difficult to find in the early stage. However, partial discharges (PD) is easy to occur in the defects and make the defect large, which may cause the failure of GIS and bring about huge economic losses to both power systems and society eventually. Therefore, it is helpful to discover the defect discharge in time and predict its development trend for the early warning of fault and taking suitable countermeasures. In this paper, ARMA model is selected to predict the development trend of partial discharge, and partial discharge experiment and three kinds of defect models are designed evaluate the prediction effect of the ARMA model. Finally, the conclusion is drawn that ARMA model can accurately predict the development trend of linear characteristic parameters, but it cannot predict that of irregular characteristic parameters of PD accurately.