Prediction of pollution flashovervoltage based on multiple characteristics of leakage current

With the accumulation of dirt and chemicals from the environment, outdoor insulator pollution flashover accidents tend to occur, which are threatening the operation safety of the power system. In this paper, a model is presented to predict the pollution flashover voltage by analyzing multiple characteristics of leakage current under AC operating voltage. By means of setting different equivalent salt deposit density(ESDD) and non-soluble deposit density(NSDD), insulators could be tested under simulated environment. Then, leakage current flowing along the surface of composite material insulators(FXBW4-10/70) can be monitored when an increasing voltage is applied. According to the time and frequency domain analysis, different parameters, i.e. the root mean square(RMS), proportion of harmonics and so on can also be measured. Finally, by adopting the theory ofprincipal component analysis(PCA) and artificial neural network(ANN), the value of pollution flashover voltage is predictedaccurately with simplifying data sets. This model modifies traditional diagnostic patterns which are only connected with singlecharacteristic and it is expected to become a multi-characteristic method for the prediction of pollution flashover voltage in thepower system.