Research On The Forecasting Method For The Surface Conditions Of The Suspension-type Insulator

In order to evaluate the degree of insulator filthy and prevent insulator's pollution flashover, at home and abroad there are many researches about the relationship between the leakage current(LC), insulator filthy and various environmental factors, but rarely comprehensive. The experiments are carried out in the entire insulation climate chamber, especially in saturated moisture conditions, and the effect of the equivalent salt deposit density(ESDD), the non-soluble deposit density(NSDD) and the relative humidity(RH) on the LC is analyzed. Results show that they have the similar nonlinear function. It can be categorized as filthy degrees. In order to predict the insulator filthy, the Particle Swarm Optimization Back Propagation(PSO-BP) of the artificial neural network(ANN) is employed to build the prediction model for ESDD and NSDD. At the same time using the data simulation is carried out for the prediction model. It shows that the optimized PSO-BP model has achieved good predicting performance.